indexed_doc en utf8 http://C:/Users/anupama/Desktop/GS311_14Aug2023/web/sites/localsite/collect/Word-PDF-Enhanced/tmp/1693990076_1/pdf01.html http://C:/Users/anupama/Desktop/GS311_14Aug2023/web/sites/localsite/collect/Word-PDF-Enhanced/tmp/1693990076_1/pdf01.html pdf01 import\pdf01.pdf url tmp\1693990076_1\pdf01.html pdf01.html pdf01.pdf pdf01.pdf PDFv2Plugin 269487 pdf01 PDF _iconpdf_ doc.pdf doc.pdf Ian H. Witten Rodger J. McNab Stefan J. Boddie David Bainbridge Greenstone: A Comprehensive Open-Source Digital Library Software System HASH1a9cea0f239f754007681b 1666853534 20221027 1693990079 20230906 HASH1a9c.dir page1.png:image/png: page2.png:image/png: page3.png:image/png: page4.png:image/png: page5.png:image/png: page6.png:image/png: page7.png:image/png: page8.png:image/png: page9.png:image/png: doc.pdf:application/pdf:

Page 1 <div id="page1"> <div style="position: relative; height: 792px; width: 612px;"><style type="text/css"> .txt { white-space:nowrap; } .p1f0 { font-family:sans-serif; font-weight:normal; font-style:normal; } .p1f1 { font-family:serif; font-weight:bold; font-style:normal; } .p1f2 { font-family:serif; font-weight:normal; font-style:italic; } .p1f3 { font-family:serif; font-weight:normal; font-style:normal; } .p1f4 { font-family:sans-serif; font-weight:bold; font-style:normal; } .p1f5 { font-family:sans-serif; font-weight:normal; font-style:normal; } </style> <img class="background" height="792" id="background1" src="_httpdocimg_/page1.png" style="position:absolute; left:0px; top:0px;" width="612"> <div class="txt" style="position:absolute; left:140px; top:60px;"><span class="p1f1" style="font-size:17px;vertical-align:baseline;color:rgba(0,0,0,1);">Greenstone: A Comprehensive Open-Source</span></div> <div class="txt" style="position:absolute; left:185px; top:80px;"><span class="p1f1" style="font-size:17px;vertical-align:baseline;color:rgba(0,0,0,1);">Digital Library Software System</span></div> <div class="txt" style="position:absolute; left:111px; top:109px;"><span class="p1f2" style="font-size:13px;vertical-align:baseline;color:rgba(0,0,0,1);">Ian H. Witten,</span><span class="p1f2" style="font-size:8px;vertical-align:super;color:rgba(0,0,0,1);">* </span><span class="p1f2" style="font-size:13px;vertical-align:baseline;color:rgba(0,0,0,1);">Rodger J. McNab,</span><span class="p1f2" style="font-size:8px;vertical-align:super;color:rgba(0,0,0,1);">† </span><span class="p1f2" style="font-size:13px;vertical-align:baseline;color:rgba(0,0,0,1);">Stefan J. Boddie,</span><span class="p1f2" style="font-size:8px;vertical-align:super;color:rgba(0,0,0,1);">* </span><span class="p1f2" style="font-size:13px;vertical-align:baseline;color:rgba(0,0,0,1);">David Bainbridge</span><span class="p1f2" style="font-size:8px;vertical-align:super;color:rgba(0,0,0,1);">*</span></div> <div class="txt" style="position:absolute; left:124px; top:129px;"><span class="p1f3" style="font-size:5px;vertical-align:baseline;color:rgba(0,0,0,1);">* </span><span class="p1f3" style="font-size:9px;vertical-align:sub;color:rgba(0,0,0,1);">Dept of Computer Science</span></div> <div class="txt" style="position:absolute; left:107px; top:144px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">University of Waikato, New Zealand</span></div> <div class="txt" style="position:absolute; left:79px; top:158px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">E-mail: {ihw, sjboddie, davidb}@cs.waikato.ac.nz</span></div> <div class="txt" style="position:absolute; left:400px; top:129px;"><span class="p1f3" style="font-size:5px;vertical-align:baseline;color:rgba(0,0,0,1);">† </span><span class="p1f3" style="font-size:9px;vertical-align:sub;color:rgba(0,0,0,1);">Digilib Systems</span></div> <div class="txt" style="position:absolute; left:387px; top:144px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Hamilton, New Zealand</span></div> <div class="txt" style="position:absolute; left:375px; top:158px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">E-mail: rodger@digilibs.com</span></div> <div class="txt" style="position:absolute; left:63px; top:198px;"><span class="p1f4" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">ABSTRACT</span></div> <div class="txt" style="position:absolute; left:63px; top:214px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">This paper describes the Greenstone digital library</span></div> <div class="txt" style="position:absolute; left:63px; top:225px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">software, a comprehensive, open-source system for the</span></div> <div class="txt" style="position:absolute; left:63px; top:236px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">construction and presentation of information collections.</span></div> <div class="txt" style="position:absolute; left:63px; top:246px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Collections built with Greenstone offer effective full-text</span></div> <div class="txt" style="position:absolute; left:63px; top:257px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">searching and metadata-based browsing facilities that are</span></div> <div class="txt" style="position:absolute; left:63px; top:267px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">attractive and easy to use. Moreover, they are easily</span></div> <div class="txt" style="position:absolute; left:63px; top:278px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">maintainable and can be augmented and rebuilt entirely</span></div> <div class="txt" style="position:absolute; left:63px; top:288px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">automatically. The system is extensible: software</span></div> <div class="txt" style="position:absolute; left:63px; top:299px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">“plugins” accommodate different document and metadata</span></div> <div class="txt" style="position:absolute; left:63px; top:309px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">types.</span></div> <div class="txt" style="position:absolute; left:63px; top:332px;"><span class="p1f4" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">INTRODUCTION</span></div> <div class="txt" style="position:absolute; left:63px; top:348px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Notwithstanding intense research activity in the digital</span></div> <div class="txt" style="position:absolute; left:63px; top:358px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">library field during the second half of the 1990s,</span></div> <div class="txt" style="position:absolute; left:63px; top:369px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">comprehensive software systems for creating digital</span></div> <div class="txt" style="position:absolute; left:63px; top:379px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">libraries are not widely available. In fact, the usual solution</span></div> <div class="txt" style="position:absolute; left:63px; top:390px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">when creating a digital library is also the most</span></div> <div class="txt" style="position:absolute; left:63px; top:400px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">obvious—just put it on the Web. But consider how much</span></div> <div class="txt" style="position:absolute; left:63px; top:411px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">effort is involved in constructing a Web site for a digital</span></div> <div class="txt" style="position:absolute; left:63px; top:422px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">library. To be effective it needs to be visually attractive</span></div> <div class="txt" style="position:absolute; left:63px; top:432px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and ergonomically easy to use, incorporate convenient and</span></div> <div class="txt" style="position:absolute; left:63px; top:443px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">powerful searching capabilities, and offer rich and natural</span></div> <div class="txt" style="position:absolute; left:63px; top:453px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">browsing facilities. Above all it must be easy to maintain</span></div> <div class="txt" style="position:absolute; left:63px; top:464px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and augment, which presents a significant challenge if any</span></div> <div class="txt" style="position:absolute; left:63px; top:474px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">manual organization is involved.</span></div> <div class="txt" style="position:absolute; left:63px; top:491px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">The alternative is to automate these activities through</span></div> <div class="txt" style="position:absolute; left:63px; top:501px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">software tools. But the broad scope of digital library</span></div> <div class="txt" style="position:absolute; left:63px; top:512px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">requirements makes this a daunting prospect. Ideally the</span></div> <div class="txt" style="position:absolute; left:63px; top:522px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">software should incorporate facilities ranging from</span></div> <div class="txt" style="position:absolute; left:320px; top:198px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">multilingual information retrieval to distributed computing</span></div> <div class="txt" style="position:absolute; left:320px; top:209px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">protocols, from interoperability to search engine</span></div> <div class="txt" style="position:absolute; left:320px; top:219px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">technology, from metadata standards to multiformat</span></div> <div class="txt" style="position:absolute; left:320px; top:230px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">document parsing, from multimedia to multiple operating</span></div> <div class="txt" style="position:absolute; left:320px; top:240px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">systems, from Web browsers to plug-and-play DVDs.</span></div> <div class="txt" style="position:absolute; left:320px; top:257px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">The Greenstone Digital Library Software from the New</span></div> <div class="txt" style="position:absolute; left:320px; top:267px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Zealand Digital Library (NZDL) project tackles this issue</span></div> <div class="txt" style="position:absolute; left:320px; top:278px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">by providing a new way of organizing information and</span></div> <div class="txt" style="position:absolute; left:320px; top:288px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">making it available over the Internet. A </span><span class="p1f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">collection </span><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">of</span></div> <div class="txt" style="position:absolute; left:320px; top:299px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">information comprises several (typically several thousand,</span></div> <div class="txt" style="position:absolute; left:320px; top:309px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">or several million) </span><span class="p1f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">documents</span><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, and a uniform interface is</span></div> <div class="txt" style="position:absolute; left:320px; top:320px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">provided to all documents in a collection. A library may</span></div> <div class="txt" style="position:absolute; left:320px; top:331px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">include many different collections, each organized</span></div> <div class="txt" style="position:absolute; left:320px; top:341px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">differently—though there is a strong family resemblance in</span></div> <div class="txt" style="position:absolute; left:320px; top:352px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">how collections are presented.</span></div> <div class="txt" style="position:absolute; left:320px; top:368px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Making information available using this system is far more</span></div> <div class="txt" style="position:absolute; left:320px; top:379px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">than “just putting it on the Web.” The collection becomes</span></div> <div class="txt" style="position:absolute; left:320px; top:389px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">maintainable, searchable, and browsable. Each collection,</span></div> <div class="txt" style="position:absolute; left:320px; top:400px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">prior to presentation, undergoes a “building” process that,</span></div> <div class="txt" style="position:absolute; left:320px; top:410px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">once established, is completely automatic. This process</span></div> <div class="txt" style="position:absolute; left:320px; top:421px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">creates all the structures that are used at run-time for</span></div> <div class="txt" style="position:absolute; left:320px; top:431px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">accessing the collection. Searching is based on various</span></div> <div class="txt" style="position:absolute; left:320px; top:442px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">indexes, while browsing is based on various metadata;</span></div> <div class="txt" style="position:absolute; left:320px; top:453px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">support structures for both are created during the building</span></div> <div class="txt" style="position:absolute; left:320px; top:463px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">operation. When new material appears it can be fully</span></div> <div class="txt" style="position:absolute; left:320px; top:474px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">incorporated into the collection by rebuilding.</span></div> <div class="txt" style="position:absolute; left:320px; top:490px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">To address the exceptionally broad demands of digital</span></div> <div class="txt" style="position:absolute; left:320px; top:501px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">libraries, the system is public and extensible. It is issued</span></div> <div class="txt" style="position:absolute; left:320px; top:511px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">under the Gnu public license and, in the spirit of open-</span></div> <div class="txt" style="position:absolute; left:320px; top:522px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">source software, users are invited to contribute</span></div> <div class="txt" style="position:absolute; left:320px; top:532px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">modifications and enhancements. Only through an</span></div> <div class="txt" style="position:absolute; left:320px; top:543px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">international cooperative effort will digital library software</span></div> <div class="txt" style="position:absolute; left:320px; top:553px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">become sufficiently comprehensive to meet the world’s</span></div> <div class="txt" style="position:absolute; left:320px; top:564px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">needs. Currently the Greenstone software is used at sites in</span></div> <div class="txt" style="position:absolute; left:320px; top:574px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Canada, Germany, New Zealand, Romania, UK, and the</span></div> <div class="txt" style="position:absolute; left:320px; top:585px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">US, and collections range from newspaper articles to</span></div> <div class="txt" style="position:absolute; left:320px; top:596px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">technical documents, from educational journals to oral</span></div> <div class="txt" style="position:absolute; left:320px; top:606px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">history, from visual art to folksongs. The software has</span></div> <div class="txt" style="position:absolute; left:320px; top:617px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">been used for collections in many different languages, and</span></div> <div class="txt" style="position:absolute; left:320px; top:627px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">for CD-ROMs that have been published by the United</span></div> <div class="txt" style="position:absolute; left:320px; top:638px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Nations and other humanitarian agencies in Belgium,</span></div> <div class="txt" style="position:absolute; left:320px; top:648px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">France, Japan, and the US for distribution in developing</span></div> <div class="txt" style="position:absolute; left:320px; top:659px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">countries (Humanity Libraries, 1998; PAHO, 1999;</span></div> <div class="txt" style="position:absolute; left:320px; top:670px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">UNESCO, 1999; UNU, 1998). Further details can be</span></div> <div class="txt" style="position:absolute; left:320px; top:680px;"><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">obtained from </span><span class="p1f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">www.nzdl.org</span><span class="p1f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">.</span></div> </div> </div>

Page 2 <div id="page2"> <div style="position: relative; height: 792px; width: 612px;"><style type="text/css"> .txt { white-space:nowrap; } .p2f0 { font-family:sans-serif; font-weight:normal; font-style:normal; } .p2f1 { font-family:serif; font-weight:normal; font-style:normal; } .p2f2 { font-family:sans-serif; font-weight:bold; font-style:normal; } .p2f3 { font-family:sans-serif; font-weight:normal; font-style:normal; } .p2f4 { font-family:serif; font-weight:normal; font-style:italic; } .p2f5 { font-family:sans-serif; font-weight:normal; font-style:normal; } </style> <img class="background" height="792" id="background2" src="_httpdocimg_/page2.png" style="position:absolute; left:0px; top:0px;" width="612"> <div class="txt" style="position:absolute; left:112px; top:264px;"><span class="p2f2" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">Figure 1: Searching the HDL collection</span></div> <div class="txt" style="position:absolute; left:63px; top:290px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">This paper sets the scene with a brief discussion of what a</span></div> <div class="txt" style="position:absolute; left:63px; top:301px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">digital library is. We then give an overview of the facilities</span></div> <div class="txt" style="position:absolute; left:63px; top:311px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">offered by Greenstone and show how end users find</span></div> <div class="txt" style="position:absolute; left:63px; top:322px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">information in collections. Next we describe the files and</span></div> <div class="txt" style="position:absolute; left:63px; top:333px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">directories involved in a collection, and then discuss the</span></div> <div class="txt" style="position:absolute; left:63px; top:343px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">processes of updating existing collections and creating new</span></div> <div class="txt" style="position:absolute; left:63px; top:354px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">ones, including extending the software to provide new</span></div> <div class="txt" style="position:absolute; left:63px; top:364px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">facilities. We conclude with an overview of related work.</span></div> <div class="txt" style="position:absolute; left:63px; top:387px;"><span class="p2f2" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">WHAT IS A DIGITAL LIBRARY?</span></div> <div class="txt" style="position:absolute; left:63px; top:403px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Ten definitions of the term “digital library” have been</span></div> <div class="txt" style="position:absolute; left:63px; top:413px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">culled from the literature by Fox (1998), and their spirit is</span></div> <div class="txt" style="position:absolute; left:63px; top:424px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">captured in the following brief characterization:</span></div> <div class="txt" style="position:absolute; left:85px; top:440px;"><span class="p2f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">A collection of digital objects, including text,</span></div> <div class="txt" style="position:absolute; left:85px; top:451px;"><span class="p2f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">video, and audio, along with methods for access</span></div> <div class="txt" style="position:absolute; left:85px; top:461px;"><span class="p2f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and retrieval, and for selection, organization</span></div> <div class="txt" style="position:absolute; left:85px; top:472px;"><span class="p2f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and maintenance of the collection</span></div> <div class="txt" style="position:absolute; left:63px; top:488px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">(Akscyn and Witten, 1998). Lesk (1998) views digital</span></div> <div class="txt" style="position:absolute; left:63px; top:499px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">libraries as “organized collections of digital information,”</span></div> <div class="txt" style="position:absolute; left:63px; top:509px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and wisely recommends that they articulate the principles</span></div> <div class="txt" style="position:absolute; left:63px; top:520px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">governing what is included and how the collection is</span></div> <div class="txt" style="position:absolute; left:63px; top:530px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">organized.</span></div> <div class="txt" style="position:absolute; left:63px; top:547px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Digital libraries are generally distinguished from the</span></div> <div class="txt" style="position:absolute; left:63px; top:557px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">World-Wide Web, the essential difference being in</span></div> <div class="txt" style="position:absolute; left:63px; top:568px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">selection and organization. But they are not generally</span></div> <div class="txt" style="position:absolute; left:63px; top:578px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">distinguished from a web </span><span class="p2f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">site</span><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">: indeed, virtually all extant</span></div> <div class="txt" style="position:absolute; left:63px; top:589px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">digital libraries manifest themselves as a web site. Hence</span></div> <div class="txt" style="position:absolute; left:63px; top:599px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the obvious question: to make a digital library, why not</span></div> <div class="txt" style="position:absolute; left:63px; top:610px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">just put the information on the Web?</span></div> <div class="txt" style="position:absolute; left:63px; top:626px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">But we make a distinction between a digital library and a</span></div> <div class="txt" style="position:absolute; left:63px; top:637px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">web site that lies at the heart of our software design: one</span></div> <div class="txt" style="position:absolute; left:63px; top:647px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">should easily be able to add new material to a library</span></div> <div class="txt" style="position:absolute; left:63px; top:658px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">without having to integrate it manually or edit its content</span></div> <div class="txt" style="position:absolute; left:63px; top:669px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">in any way. Once added, new material should immediately</span></div> <div class="txt" style="position:absolute; left:320px; top:53px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">become a first-class component of the library. And what</span></div> <div class="txt" style="position:absolute; left:320px; top:64px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">permits it to be integrated into existing searching and</span></div> <div class="txt" style="position:absolute; left:320px; top:74px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">browsing structures without any manual intervention is</span></div> <div class="txt" style="position:absolute; left:320px; top:85px;"><span class="p2f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">metadata</span><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">. This provides sufficient focus to the concept of</span></div> <div class="txt" style="position:absolute; left:320px; top:95px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">“digital library” to support the development of a</span></div> <div class="txt" style="position:absolute; left:320px; top:106px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">construction kit.</span></div> <div class="txt" style="position:absolute; left:320px; top:128px;"><span class="p2f2" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">OVERVIEW OF GREENSTONE</span></div> <div class="txt" style="position:absolute; left:320px; top:144px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Information collections built by Greenstone combine</span></div> <div class="txt" style="position:absolute; left:320px; top:155px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">extensive full-text search facilities with browsing indexes</span></div> <div class="txt" style="position:absolute; left:320px; top:165px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">based on different metadata types. There are several ways</span></div> <div class="txt" style="position:absolute; left:320px; top:176px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">for users to find information, although they differ between</span></div> <div class="txt" style="position:absolute; left:320px; top:187px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">collections depending on the metadata available and the</span></div> <div class="txt" style="position:absolute; left:320px; top:197px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">collection design. Typically you can </span><span class="p2f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">search for particular</span></div> <div class="txt" style="position:absolute; left:320px; top:208px;"><span class="p2f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">words </span><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">that appear in the text, or within a section of a</span></div> <div class="txt" style="position:absolute; left:320px; top:218px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">document, or within a title or section heading. You can</span></div> <div class="txt" style="position:absolute; left:320px; top:229px;"><span class="p2f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">browse documents by title</span><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">: just click on the displayed book</span></div> <div class="txt" style="position:absolute; left:320px; top:239px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">icon to read it. You can </span><span class="p2f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">browse documents by subject</span><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">.</span></div> <div class="txt" style="position:absolute; left:320px; top:250px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Subjects are represented by bookshelves: just click on a</span></div> <div class="txt" style="position:absolute; left:320px; top:261px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">shelf to see the books. Where appropriate, documents</span></div> <div class="txt" style="position:absolute; left:320px; top:271px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">come complete with a table of contents (constructed</span></div> <div class="txt" style="position:absolute; left:320px; top:282px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">automatically): you can click on a chapter or subsection to</span></div> <div class="txt" style="position:absolute; left:320px; top:292px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">open it, expand the full table of contents, or expand the full</span></div> <div class="txt" style="position:absolute; left:320px; top:303px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">document.</span></div> <div class="txt" style="position:absolute; left:320px; top:319px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">An example of searching is shown in Figure 1 where</span></div> <div class="txt" style="position:absolute; left:320px; top:330px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">documents in the Global Help Project’s Humanity</span></div> <div class="txt" style="position:absolute; left:320px; top:340px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Development Library (HDL) are being searched for</span></div> <div class="txt" style="position:absolute; left:320px; top:351px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">chapters matching the word </span><span class="p2f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">butterfly</span><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">. In Figure 2 the same</span></div> <div class="txt" style="position:absolute; left:320px; top:361px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">collection is being browsed by subject: by clicking on the</span></div> <div class="txt" style="position:absolute; left:320px; top:372px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">bookshelf icons the user has discovered an item under</span></div> <div class="txt" style="position:absolute; left:320px; top:382px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Section 16, Animal Husbandry. Pursuing an interest in</span></div> <div class="txt" style="position:absolute; left:320px; top:393px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">butterfly farming, the user selects a book by clicking on its</span></div> <div class="txt" style="position:absolute; left:320px; top:404px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">book icon. In Figure 3 the front cover of the book is</span></div> <div class="txt" style="position:absolute; left:320px; top:414px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">displayed as a graphic on the left, and the automatically</span></div> <div class="txt" style="position:absolute; left:320px; top:425px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">constructed table of contents appears at the start of the</span></div> <div class="txt" style="position:absolute; left:320px; top:435px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">document. The current focus, </span><span class="p2f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Introduction and Summary</span><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">,</span></div> <div class="txt" style="position:absolute; left:320px; top:446px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">is shown in bold in the table of contents with its text</span></div> <div class="txt" style="position:absolute; left:320px; top:456px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">starting further down the page.</span></div> <div class="txt" style="position:absolute; left:320px; top:473px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">In accordance with Lesk’s advice, a statement of purpose</span></div> <div class="txt" style="position:absolute; left:320px; top:483px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and coverage accompanies each collection, along with an</span></div> <div class="txt" style="position:absolute; left:320px; top:494px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">explanation of how it is organized (Figure 1 shows the</span></div> <div class="txt" style="position:absolute; left:320px; top:504px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">start of this). A distinction is made between </span><span class="p2f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">searching </span><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and</span></div> <div class="txt" style="position:absolute; left:320px; top:515px;"><span class="p2f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">browsing</span><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">. Searching is full-text, and—depending on the</span></div> <div class="txt" style="position:absolute; left:320px; top:525px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">collection’s design—the user can choose between indexes</span></div> <div class="txt" style="position:absolute; left:320px; top:536px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">built from different parts of the documents, or from</span></div> <div class="txt" style="position:absolute; left:320px; top:547px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">different metadata. Some collections have an index of full</span></div> <div class="txt" style="position:absolute; left:320px; top:557px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">documents, an index of sections, an index of paragraphs,</span></div> <div class="txt" style="position:absolute; left:320px; top:568px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">an index of titles, and an index of section headings, each of</span></div> <div class="txt" style="position:absolute; left:320px; top:578px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">which can be searched for particular words or phrases.</span></div> <div class="txt" style="position:absolute; left:320px; top:589px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Browsing involves data structures created from metadata</span></div> <div class="txt" style="position:absolute; left:320px; top:599px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">that the user can examine: lists of authors, lists of titles,</span></div> <div class="txt" style="position:absolute; left:320px; top:610px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">lists of dates, hierarchical classification structures, and so</span></div> <div class="txt" style="position:absolute; left:320px; top:621px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">on. Data structures for both browsing and searching are</span></div> <div class="txt" style="position:absolute; left:320px; top:631px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">built according to instructions in a configuration file,</span></div> <div class="txt" style="position:absolute; left:320px; top:642px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">which controls both building and serving the collection.</span></div> <div class="txt" style="position:absolute; left:320px; top:652px;"><span class="p2f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Sample configuration files are discussed below.</span></div> </div> </div>

Page 3 <div id="page3"> <div style="position: relative; height: 792px; width: 612px;"><style type="text/css"> .txt { white-space:nowrap; } .p3f0 { font-family:sans-serif; font-weight:normal; font-style:normal; } .p3f1 { font-family:sans-serif; font-weight:normal; font-style:normal; } .p3f2 { font-family:serif; font-weight:normal; font-style:normal; } .p3f3 { font-family:serif; font-weight:normal; font-style:italic; } .p3f4 { font-family:sans-serif; font-weight:bold; font-style:normal; } .p3f5 { font-family:sans-serif; font-weight:normal; font-style:normal; } </style> <img class="background" height="792" id="background3" src="_httpdocimg_/page3.png" style="position:absolute; left:0px; top:0px;" width="612"> <div class="txt" style="position:absolute; left:96px; top:264px;"><span class="p3f4" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">Figure 2: Browsing the HDL collection by subject</span></div> <div class="txt" style="position:absolute; left:63px; top:290px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Rich browsing facilities can be provided by manually</span></div> <div class="txt" style="position:absolute; left:63px; top:301px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">linking parts of documents together and building explicit</span></div> <div class="txt" style="position:absolute; left:63px; top:311px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">indexes and tables of contents. However, manually-created</span></div> <div class="txt" style="position:absolute; left:63px; top:322px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">linking becomes difficult to maintain, and often falls into</span></div> <div class="txt" style="position:absolute; left:63px; top:333px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">disrepair when a collection expands. The Greenstone</span></div> <div class="txt" style="position:absolute; left:63px; top:343px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">software takes a different tack: it facilitates </span><span class="p3f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">maintainability</span></div> <div class="txt" style="position:absolute; left:63px; top:354px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">by creating all searching and browsing structures</span></div> <div class="txt" style="position:absolute; left:63px; top:364px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">automatically from the documents themselves. No links</span></div> <div class="txt" style="position:absolute; left:63px; top:375px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">are inserted by hand. This means that when new</span></div> <div class="txt" style="position:absolute; left:63px; top:385px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">documents in the same format become available, they can</span></div> <div class="txt" style="position:absolute; left:63px; top:396px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">be added automatically. Indeed, for some collections this is</span></div> <div class="txt" style="position:absolute; left:63px; top:406px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">done by processes that wake up regularly, scout for new</span></div> <div class="txt" style="position:absolute; left:63px; top:417px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">material, and rebuild the indexes—all without manual</span></div> <div class="txt" style="position:absolute; left:63px; top:428px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">intervention.</span></div> <div class="txt" style="position:absolute; left:63px; top:444px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Collections comprise many documents: thousands, tens of</span></div> <div class="txt" style="position:absolute; left:63px; top:454px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">thousands, or even millions. Each document may be</span></div> <div class="txt" style="position:absolute; left:63px; top:465px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">hierarchically organized into </span><span class="p3f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">sections </span><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">(subsections, sub-</span></div> <div class="txt" style="position:absolute; left:63px; top:476px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">subsections, and so on). Each section comprises one or</span></div> <div class="txt" style="position:absolute; left:63px; top:486px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">more </span><span class="p3f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">paragraphs</span><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">. Metadata such as author, title, date,</span></div> <div class="txt" style="position:absolute; left:63px; top:497px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">keywords, and so on, may be associated with documents,</span></div> <div class="txt" style="position:absolute; left:63px; top:507px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">or with individual sections of documents. This is the raw</span></div> <div class="txt" style="position:absolute; left:63px; top:518px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">material for indexes. It must either be provided explicitly</span></div> <div class="txt" style="position:absolute; left:63px; top:528px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">for each document and section (for example, in an</span></div> <div class="txt" style="position:absolute; left:63px; top:539px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">accompanying spreadsheet) or be derivable automatically</span></div> <div class="txt" style="position:absolute; left:63px; top:549px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">from the source documents. Metadata is converted to</span></div> <div class="txt" style="position:absolute; left:63px; top:560px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Dublin Core and stored with the document for internal use.</span></div> <div class="txt" style="position:absolute; left:63px; top:576px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">In order to accommodate different kinds of source</span></div> <div class="txt" style="position:absolute; left:63px; top:587px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">documents, the software is organized so that “plugins” can</span></div> <div class="txt" style="position:absolute; left:63px; top:597px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">be written for new document types. Plugins exist for plain</span></div> <div class="txt" style="position:absolute; left:63px; top:608px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">text documents, HTML documents, email documents, and</span></div> <div class="txt" style="position:absolute; left:63px; top:619px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">bibliographic formats. Word documents are handled by</span></div> <div class="txt" style="position:absolute; left:63px; top:629px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">saving them as HTML; PostScript ones by applying a</span></div> <div class="txt" style="position:absolute; left:63px; top:640px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">preprocessor (Nevill-Manning </span><span class="p3f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">et al</span><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">., 1998). Specially</span></div> <div class="txt" style="position:absolute; left:63px; top:650px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">written plugins also exist for proprietary formats such as</span></div> <div class="txt" style="position:absolute; left:63px; top:661px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">that used by the BBC archives department. A collection</span></div> <div class="txt" style="position:absolute; left:63px; top:671px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">may have source documents in different forms: it is just a</span></div> <div class="txt" style="position:absolute; left:320px; top:53px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">matter of specifying all the necessary plugins. In order to</span></div> <div class="txt" style="position:absolute; left:320px; top:64px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">build browsing indexes from metadata, an analogous</span></div> <div class="txt" style="position:absolute; left:320px; top:74px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">scheme of “classifiers” is used: classifiers create indexes</span></div> <div class="txt" style="position:absolute; left:320px; top:85px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">of various kinds based on metadata. Source documents are</span></div> <div class="txt" style="position:absolute; left:320px; top:95px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">brought into the Greenstone system through a process</span></div> <div class="txt" style="position:absolute; left:320px; top:106px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">called </span><span class="p3f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">importing</span><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, which uses the plugins and classifiers</span></div> <div class="txt" style="position:absolute; left:320px; top:117px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">specified in the collection configuration file.</span></div> <div class="txt" style="position:absolute; left:320px; top:133px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">The international Unicode character set is used throughout,</span></div> <div class="txt" style="position:absolute; left:320px; top:143px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">so documents—and interfaces—can be written in any</span></div> <div class="txt" style="position:absolute; left:320px; top:154px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">language. Collections have so far been produced in</span></div> <div class="txt" style="position:absolute; left:320px; top:165px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">English, French, Spanish, German, Maori, Chinese, and</span></div> <div class="txt" style="position:absolute; left:320px; top:175px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Arabic. The NZDL Web site provides numerous examples.</span></div> <div class="txt" style="position:absolute; left:320px; top:186px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Collections can contain text, pictures, and even audio and</span></div> <div class="txt" style="position:absolute; left:320px; top:196px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">video clips; a text-only version of the interface is also</span></div> <div class="txt" style="position:absolute; left:320px; top:207px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">provided to accommodate visually impaired users.</span></div> <div class="txt" style="position:absolute; left:320px; top:217px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Compression technology is used to ensure best use of</span></div> <div class="txt" style="position:absolute; left:320px; top:228px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">storage (Witten </span><span class="p3f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">et al </span><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">., 1999). Most non-textual material is</span></div> <div class="txt" style="position:absolute; left:320px; top:238px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">either linked to textual documents or accompanied by</span></div> <div class="txt" style="position:absolute; left:320px; top:249px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">textual descriptions (such as photo captions) to allow full-</span></div> <div class="txt" style="position:absolute; left:320px; top:260px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">text searching and browsing. However, the architecture</span></div> <div class="txt" style="position:absolute; left:320px; top:270px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">permits the implementation of plugins and classifiers even</span></div> <div class="txt" style="position:absolute; left:320px; top:281px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">for non-textual data.</span></div> <div class="txt" style="position:absolute; left:320px; top:297px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">The system includes an “administrative” function whereby</span></div> <div class="txt" style="position:absolute; left:320px; top:308px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">specified users can examine the composition of all</span></div> <div class="txt" style="position:absolute; left:320px; top:318px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">collections, protect documents so that they can only be</span></div> <div class="txt" style="position:absolute; left:320px; top:329px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">accessed by registered users on presentation of a password,</span></div> <div class="txt" style="position:absolute; left:320px; top:339px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and so on. Logs of user activity are kept that record all</span></div> <div class="txt" style="position:absolute; left:320px; top:350px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">queries made to every Greenstone collection (though this</span></div> <div class="txt" style="position:absolute; left:320px; top:360px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">facility can be disabled).</span></div> <div class="txt" style="position:absolute; left:320px; top:377px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Although primarily designed for Internet access over the</span></div> <div class="txt" style="position:absolute; left:320px; top:387px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">World-Wide Web, collections can be made available, in</span></div> <div class="txt" style="position:absolute; left:320px; top:398px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">precisely the same form, on CD-ROM. In either case they</span></div> <div class="txt" style="position:absolute; left:320px; top:408px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">are accessed through any Web browser. Greenstone CD-</span></div> <div class="txt" style="position:absolute; left:320px; top:419px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">ROMs operate on a standalone PC under Windows 3.X,</span></div> <div class="txt" style="position:absolute; left:320px; top:429px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">95, 98, and NT, and the interaction is identical to accessing</span></div> <div class="txt" style="position:absolute; left:320px; top:440px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the collection on the Web—except that response is faster</span></div> <div class="txt" style="position:absolute; left:320px; top:451px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and more predictable. The requirement to operate on early</span></div> <div class="txt" style="position:absolute; left:320px; top:461px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Windows systems is one that plagues the software design,</span></div> <div class="txt" style="position:absolute; left:320px; top:472px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">but is crucial for many users—particularly those in</span></div> <div class="txt" style="position:absolute; left:320px; top:482px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">underdeveloped countries seeking access to humanitarian</span></div> <div class="txt" style="position:absolute; left:320px; top:493px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">aid collections. If the PC is connected to a network</span></div> <div class="txt" style="position:absolute; left:320px; top:503px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">(intranet or Internet), a custom-built Web server provided</span></div> <div class="txt" style="position:absolute; left:320px; top:514px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">on each CD makes exactly the same information available</span></div> <div class="txt" style="position:absolute; left:320px; top:525px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">to others through their standard Web browser. The use of</span></div> <div class="txt" style="position:absolute; left:320px; top:535px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">compression ensures that the greatest possible volume of</span></div> <div class="txt" style="position:absolute; left:320px; top:546px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">information can be packed on to a CD-ROM.</span></div> <div class="txt" style="position:absolute; left:320px; top:562px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">The collection-serving software operates under Unix and</span></div> <div class="txt" style="position:absolute; left:320px; top:573px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Windows NT, and works with standard Web servers. A</span></div> <div class="txt" style="position:absolute; left:320px; top:583px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">flexible process structure allows different collections to be</span></div> <div class="txt" style="position:absolute; left:320px; top:594px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">served by different computers, yet be presented to the user</span></div> <div class="txt" style="position:absolute; left:320px; top:604px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">in the same way, on the same Web page, as part of the</span></div> <div class="txt" style="position:absolute; left:320px; top:615px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">same digital library, even as part of the same collection</span></div> <div class="txt" style="position:absolute; left:320px; top:625px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">(McNab and Witten, 1998). Existing collections can be</span></div> <div class="txt" style="position:absolute; left:320px; top:636px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">updated and new ones brought on-line at any time, without</span></div> <div class="txt" style="position:absolute; left:320px; top:646px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">bringing the system down; the process responsible for the</span></div> <div class="txt" style="position:absolute; left:320px; top:657px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">user interface will notice (through periodic polling) when</span></div> <div class="txt" style="position:absolute; left:320px; top:668px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">new collections appear and add them to the list presented</span></div> <div class="txt" style="position:absolute; left:320px; top:678px;"><span class="p3f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">to the user.</span></div> </div> </div>

Page 4 <div id="page4"> <div style="position: relative; height: 792px; width: 612px;"><style type="text/css"> .txt { white-space:nowrap; } .p4f0 { font-family:sans-serif; font-weight:normal; font-style:normal; } .p4f1 { font-family:sans-serif; font-weight:bold; font-style:normal; } .p4f2 { font-family:sans-serif; font-weight:normal; font-style:normal; } .p4f3 { font-family:serif; font-weight:normal; font-style:normal; } .p4f4 { font-family:serif; font-weight:normal; font-style:italic; } .p4f5 { font-family:sans-serif; font-weight:normal; font-style:normal; } </style> <img class="background" height="792" id="background4" src="_httpdocimg_/page4.png" style="position:absolute; left:0px; top:0px;" width="612"> <div class="txt" style="position:absolute; left:114px; top:261px;"><span class="p4f1" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">Figure 3: Reading a book in the HDL</span></div> <div class="txt" style="position:absolute; left:63px; top:291px;"><span class="p4f1" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">FINDING INFORMATION</span></div> <div class="txt" style="position:absolute; left:63px; top:307px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Greenstone digital library systems generally include</span></div> <div class="txt" style="position:absolute; left:63px; top:317px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">several separate collections. A home page allows you to</span></div> <div class="txt" style="position:absolute; left:63px; top:328px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">select a collection; in addition, each collection has its own</span></div> <div class="txt" style="position:absolute; left:63px; top:339px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">“about” page that gives you information about how the</span></div> <div class="txt" style="position:absolute; left:63px; top:349px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">collection is organized and the principles governing what</span></div> <div class="txt" style="position:absolute; left:63px; top:360px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">is included.</span></div> <div class="txt" style="position:absolute; left:63px; top:376px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">All icons in the screenshots of Figures 1–4 are clickable.</span></div> <div class="txt" style="position:absolute; left:63px; top:387px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Those icons at the top of the page return to the home page,</span></div> <div class="txt" style="position:absolute; left:63px; top:397px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">provide help text, and allow you to set user interface and</span></div> <div class="txt" style="position:absolute; left:63px; top:408px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">searching preferences. The navigation bar underneath</span></div> <div class="txt" style="position:absolute; left:63px; top:418px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">gives access to the searching and browsing facilities,</span></div> <div class="txt" style="position:absolute; left:63px; top:429px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">which differ from one collection to another.</span></div> <div class="txt" style="position:absolute; left:63px; top:445px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Each of the five buttons provides a different way to find</span></div> <div class="txt" style="position:absolute; left:63px; top:456px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">information. You can </span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">search for particular words </span><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">that</span></div> <div class="txt" style="position:absolute; left:63px; top:466px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">appear in the text from the “search” page (or from the</span></div> <div class="txt" style="position:absolute; left:63px; top:477px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">“about” page of Figure 1). This collection contains indexes</span></div> <div class="txt" style="position:absolute; left:63px; top:487px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">of chapters, section titles, and entire books. The default</span></div> <div class="txt" style="position:absolute; left:63px; top:498px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">search interface is a simple one, suitable for casual users;</span></div> <div class="txt" style="position:absolute; left:63px; top:508px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">advanced searching—which allows full Boolean</span></div> <div class="txt" style="position:absolute; left:63px; top:519px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">expressions, phrase searching, case and stemming</span></div> <div class="txt" style="position:absolute; left:63px; top:530px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">control—can be enabled from the </span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Preferences </span><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">page.</span></div> <div class="txt" style="position:absolute; left:63px; top:546px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">This collection has four browsable metadata indexes. You</span></div> <div class="txt" style="position:absolute; left:63px; top:556px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">can </span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">access publications by subject </span><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">by clicking the </span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">subjects</span></div> <div class="txt" style="position:absolute; left:63px; top:567px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">button, which brings up a list of subjects, represented by</span></div> <div class="txt" style="position:absolute; left:63px; top:578px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">bookshelves (Figure 2). You can </span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">access publications by</span></div> <div class="txt" style="position:absolute; left:63px; top:588px;"><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">title </span><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">by clicking </span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">titles a-z </span><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">(Figure 4), which brings up a list</span></div> <div class="txt" style="position:absolute; left:63px; top:599px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">of books in alphabetic order. You can </span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">access publications</span></div> <div class="txt" style="position:absolute; left:63px; top:609px;"><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">by organization </span><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">(i.e. Dublin Core “publisher”), bringing up</span></div> <div class="txt" style="position:absolute; left:63px; top:620px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">a list of organizations. You can </span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">access publications by</span></div> <div class="txt" style="position:absolute; left:63px; top:630px;"><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">“how to” listing</span><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, yielding a list of hints defined by the</span></div> <div class="txt" style="position:absolute; left:63px; top:641px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">collection’s editors. We use the Dublin Core as a base and</span></div> <div class="txt" style="position:absolute; left:63px; top:652px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">extend it in an </span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">ad hoc </span><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">manner to accommodate the</span></div> <div class="txt" style="position:absolute; left:63px; top:662px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">individual requirements of collection designers.</span></div> <div class="txt" style="position:absolute; left:320px; top:54px;"><span class="p4f1" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">FILES IN A COLLECTION</span></div> <div class="txt" style="position:absolute; left:320px; top:69px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">When a new collection is created or material is added to an</span></div> <div class="txt" style="position:absolute; left:320px; top:80px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">existing one, the original source documents are first</span></div> <div class="txt" style="position:absolute; left:320px; top:91px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">brought into the system through a process known as</span></div> <div class="txt" style="position:absolute; left:320px; top:101px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">“importing.” This involves converting documents into a</span></div> <div class="txt" style="position:absolute; left:320px; top:112px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">simple HTML-like format known as GML (for</span></div> <div class="txt" style="position:absolute; left:320px; top:122px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">“Greenstone Markup Language”), which includes any</span></div> <div class="txt" style="position:absolute; left:320px; top:133px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">metadata associated with the document. Documents are</span></div> <div class="txt" style="position:absolute; left:320px; top:143px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">assumed to be in the Unicode UTF-8 code (of which the</span></div> <div class="txt" style="position:absolute; left:320px; top:154px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">ASCII characters form a subset).</span></div> <div class="txt" style="position:absolute; left:320px; top:184px;"><span class="p4f1" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">Files and directories</span></div> <div class="txt" style="position:absolute; left:320px; top:200px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">There is a separate directory for each collection, which</span></div> <div class="txt" style="position:absolute; left:320px; top:210px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">contains five subdirectories: the original raw material</span></div> <div class="txt" style="position:absolute; left:320px; top:221px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">(</span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">import</span><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">), the GML files created from this (</span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">archives</span><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">), the</span></div> <div class="txt" style="position:absolute; left:320px; top:231px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">final collection as it is served to users (</span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">index</span><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">), a directory</span></div> <div class="txt" style="position:absolute; left:320px; top:242px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">for use during the building process (</span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">building</span><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">), and one for</span></div> <div class="txt" style="position:absolute; left:320px; top:253px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">any supporting files (</span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">etc</span><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">)—including the configuration file</span></div> <div class="txt" style="position:absolute; left:320px; top:263px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">that controls the collection creation procedure. Additional</span></div> <div class="txt" style="position:absolute; left:320px; top:274px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">files might be required: for example, building a hierarchy</span></div> <div class="txt" style="position:absolute; left:320px; top:284px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">of classifications requires a data file of sub-classifications.</span></div> <div class="txt" style="position:absolute; left:320px; top:314px;"><span class="p4f1" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">The imported documents</span></div> <div class="txt" style="position:absolute; left:320px; top:330px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">In order to identify documents internally, a unique object</span></div> <div class="txt" style="position:absolute; left:320px; top:341px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">identifier or OID is assigned to each original source</span></div> <div class="txt" style="position:absolute; left:320px; top:351px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">document when it is imported (formed by hashing the</span></div> <div class="txt" style="position:absolute; left:320px; top:362px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">content, to overcome file duplication effects caused by</span></div> <div class="txt" style="position:absolute; left:320px; top:372px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">mirroring) and stored as metadata within that document. It</span></div> <div class="txt" style="position:absolute; left:320px; top:383px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">is important that OIDs persist throughout the index-</span></div> <div class="txt" style="position:absolute; left:320px; top:393px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">building process—so that a user’s search history is</span></div> <div class="txt" style="position:absolute; left:320px; top:404px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">unaffected by rebuilding the collection. OIDs are assigned</span></div> <div class="txt" style="position:absolute; left:320px; top:415px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">by hashing the contents of the original source document.</span></div> <div class="txt" style="position:absolute; left:320px; top:431px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Once imported, each document is stored in its own</span></div> <div class="txt" style="position:absolute; left:320px; top:441px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">subdirectory of </span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">archives</span><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, along with any associated</span></div> <div class="txt" style="position:absolute; left:320px; top:452px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">files—for example, images. To ensure compatibility with</span></div> <div class="txt" style="position:absolute; left:320px; top:463px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Windows 3.0, only eight characters are used in directory</span></div> <div class="txt" style="position:absolute; left:320px; top:473px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and file names, which causes annoying but essentially</span></div> <div class="txt" style="position:absolute; left:320px; top:484px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">trivial complications.</span></div> <div class="txt" style="position:absolute; left:320px; top:514px;"><span class="p4f1" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">Inside the documents</span></div> <div class="txt" style="position:absolute; left:320px; top:530px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">The GML format imposes a limited amount of structure on</span></div> <div class="txt" style="position:absolute; left:320px; top:540px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">documents. Documents are divided into paragraphs. They</span></div> <div class="txt" style="position:absolute; left:320px; top:551px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">can be split hierarchically into sections and subsections.</span></div> <div class="txt" style="position:absolute; left:320px; top:561px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">OIDs are extended to identify these components by</span></div> <div class="txt" style="position:absolute; left:320px; top:572px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">appending numbers, separated by periods, to a document’s</span></div> <div class="txt" style="position:absolute; left:320px; top:582px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">OID. When a book is read, its section hierarchy is visible</span></div> <div class="txt" style="position:absolute; left:320px; top:593px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">as the table of contents (Figure 3). Chapters, sections,</span></div> <div class="txt" style="position:absolute; left:320px; top:603px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">subsections, and pages are all implemented simply as</span></div> <div class="txt" style="position:absolute; left:320px; top:614px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">“sections” within the document. In some collections</span></div> <div class="txt" style="position:absolute; left:320px; top:625px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">documents do not have a hierarchical subsection structure,</span></div> <div class="txt" style="position:absolute; left:320px; top:635px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">but are split into pages to permit browsing within a</span></div> <div class="txt" style="position:absolute; left:320px; top:646px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">retrieved document.</span></div> <div class="txt" style="position:absolute; left:320px; top:662px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">The document structure is used for searchable indexes.</span></div> <div class="txt" style="position:absolute; left:320px; top:673px;"><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">There are three levels of index: </span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">documents</span><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, </span><span class="p4f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">sections</span><span class="p4f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, and</span></div> </div> </div>

Page 5 <div id="page5"> <div style="position: relative; height: 792px; width: 612px;"><style type="text/css"> .txt { white-space:nowrap; } .p5f0 { font-family:sans-serif; font-weight:normal; font-style:normal; } .p5f1 { font-family:serif; font-weight:normal; font-style:italic; } .p5f2 { font-family:serif; font-weight:normal; font-style:normal; } .p5f3 { font-family:sans-serif; font-weight:bold; font-style:normal; } .p5f4 { font-family:sans-serif; font-weight:normal; font-style:normal; } .p5f5 { font-family:sans-serif; font-weight:normal; font-style:normal; } </style> <img class="background" height="792" id="background5" src="_httpdocimg_/page5.png" style="position:absolute; left:0px; top:0px;" width="612"> <div class="txt" style="position:absolute; left:115px; top:286px;"><span class="p5f3" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">Figure 4: Browsing titles in the HDL</span></div> <div class="txt" style="position:absolute; left:63px; top:312px;"><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">paragraphs</span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, corresponding to the distinctions that GML</span></div> <div class="txt" style="position:absolute; left:63px; top:323px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">makes—the hierarchical structure is flattened for the</span></div> <div class="txt" style="position:absolute; left:63px; top:333px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">purposes of creating these indexes. Indexes can be of text,</span></div> <div class="txt" style="position:absolute; left:63px; top:344px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">or metadata, or any combination. Thus you can create a</span></div> <div class="txt" style="position:absolute; left:63px; top:355px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">searchable index of section titles, and/or authors, and/or</span></div> <div class="txt" style="position:absolute; left:63px; top:365px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">document descriptions, as well as the document text.</span></div> <div class="txt" style="position:absolute; left:63px; top:388px;"><span class="p5f3" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">UPDATING EXISTING COLLECTIONS</span></div> <div class="txt" style="position:absolute; left:63px; top:404px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Updating an existing collection with new files in the same</span></div> <div class="txt" style="position:absolute; left:63px; top:414px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">format is easy. For example, the raw material for the HDL</span></div> <div class="txt" style="position:absolute; left:63px; top:425px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">is supplied in the form of HTML files marked up with</span></div> <div class="txt" style="position:absolute; left:63px; top:435px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">&lt;&lt;TOC&gt;&gt; tags to split books into sections and</span></div> <div class="txt" style="position:absolute; left:63px; top:446px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">subsections, and &lt;&lt;I&gt;&gt; tags to indicate where an image is</span></div> <div class="txt" style="position:absolute; left:63px; top:456px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">to be inserted. For each book in the library there is a</span></div> <div class="txt" style="position:absolute; left:63px; top:467px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">directory that contains a single HTML file representing the</span></div> <div class="txt" style="position:absolute; left:63px; top:477px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">book, and separate files containing the associated images.</span></div> <div class="txt" style="position:absolute; left:63px; top:488px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">An accompanying spreadsheet file contains the</span></div> <div class="txt" style="position:absolute; left:63px; top:499px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">classification hierarchy; this is converted to a simple file</span></div> <div class="txt" style="position:absolute; left:63px; top:509px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">format (using Excel’s </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Save As </span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">command).</span></div> <div class="txt" style="position:absolute; left:63px; top:525px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Since the collection exists, its directory is already set up</span></div> <div class="txt" style="position:absolute; left:63px; top:536px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">with subdirectories </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">import</span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">archives</span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">building</span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">index</span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, and</span></div> <div class="txt" style="position:absolute; left:63px; top:547px;"><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">etc</span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, and the </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">etc </span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">directory will contain a suitable collection</span></div> <div class="txt" style="position:absolute; left:63px; top:557px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">configuration file.</span></div> <div class="txt" style="position:absolute; left:63px; top:587px;"><span class="p5f3" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">The updating procedure</span></div> <div class="txt" style="position:absolute; left:63px; top:603px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">To update a collection, the new raw material is placed in</span></div> <div class="txt" style="position:absolute; left:63px; top:614px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">import </span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">directory, in whatever form it is available. Then</span></div> <div class="txt" style="position:absolute; left:320px; top:53px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">import </span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">process is invoked, which converts the files into</span></div> <div class="txt" style="position:absolute; left:320px; top:64px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">GML using the specified plugins. Old material for which</span></div> <div class="txt" style="position:absolute; left:320px; top:74px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">GML files have previously been created is not re-imported.</span></div> <div class="txt" style="position:absolute; left:320px; top:85px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Then the </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">build </span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">process is invoked to build the requisite</span></div> <div class="txt" style="position:absolute; left:320px; top:95px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">indexes for the collection. Finally, the contents of the</span></div> <div class="txt" style="position:absolute; left:320px; top:106px;"><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">building </span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">directory are moved into the </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">index </span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">directory, and</span></div> <div class="txt" style="position:absolute; left:320px; top:117px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the new version of the collection automatically becomes</span></div> <div class="txt" style="position:absolute; left:320px; top:127px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">live.</span></div> <div class="txt" style="position:absolute; left:320px; top:143px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">This procedure may seem cumbersome. But all the steps</span></div> <div class="txt" style="position:absolute; left:320px; top:154px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">are necessary for efficient operation with large collections.</span></div> <div class="txt" style="position:absolute; left:320px; top:165px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">The </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">import </span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">process could be performed on the fly during</span></div> <div class="txt" style="position:absolute; left:320px; top:175px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the building operation—but because building indexes is a</span></div> <div class="txt" style="position:absolute; left:320px; top:186px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">multipass operation, the often lengthy importing would be</span></div> <div class="txt" style="position:absolute; left:320px; top:196px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">repeated several times. The </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">build </span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">process can take</span></div> <div class="txt" style="position:absolute; left:320px; top:207px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">considerable time—a day or two, for very large</span></div> <div class="txt" style="position:absolute; left:320px; top:217px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">collections. Consequently, the results are placed in the</span></div> <div class="txt" style="position:absolute; left:320px; top:228px;"><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">building </span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">directory so that, if the collection already exists, it</span></div> <div class="txt" style="position:absolute; left:320px; top:238px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">will continue to be served to users in its old form</span></div> <div class="txt" style="position:absolute; left:320px; top:249px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">throughout the building operation.</span></div> <div class="txt" style="position:absolute; left:320px; top:265px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Active users of the collection will not be disturbed when</span></div> <div class="txt" style="position:absolute; left:320px; top:276px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the new version becomes live—they will probably not</span></div> <div class="txt" style="position:absolute; left:320px; top:286px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">even notice. The persistent OIDs ensure that interactions</span></div> <div class="txt" style="position:absolute; left:320px; top:297px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">remain coherent—users who are examining the results of a</span></div> <div class="txt" style="position:absolute; left:320px; top:308px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">query or browse operation will still retrieve the expected</span></div> <div class="txt" style="position:absolute; left:320px; top:318px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">documents—and if a search is actually in progress when</span></div> <div class="txt" style="position:absolute; left:320px; top:329px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the change takes place the program detects the resulting</span></div> <div class="txt" style="position:absolute; left:320px; top:339px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">file-structure inconsistency and automatically and</span></div> <div class="txt" style="position:absolute; left:320px; top:350px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">transparently re-executes the query, this time on the new</span></div> <div class="txt" style="position:absolute; left:320px; top:360px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">version of the collection.</span></div> <div class="txt" style="position:absolute; left:320px; top:390px;"><span class="p5f3" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">How it works</span></div> <div class="txt" style="position:absolute; left:320px; top:406px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">The original material in the </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">import </span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">directory may be in any</span></div> <div class="txt" style="position:absolute; left:320px; top:417px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">format, and plugins are required to process each format</span></div> <div class="txt" style="position:absolute; left:320px; top:428px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">type. The plugins that a collection uses must be specified</span></div> <div class="txt" style="position:absolute; left:320px; top:438px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">in the collection configuration file. The </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">import </span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">program</span></div> <div class="txt" style="position:absolute; left:320px; top:449px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">reads the list of plugins and passes each document to each</span></div> <div class="txt" style="position:absolute; left:320px; top:459px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">plugin in order until it finds one that can process it. When</span></div> <div class="txt" style="position:absolute; left:320px; top:470px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">updating an existing collection, all plugins necessary to</span></div> <div class="txt" style="position:absolute; left:320px; top:480px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">process new material should already have been specified in</span></div> <div class="txt" style="position:absolute; left:320px; top:491px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the configuration file.</span></div> <div class="txt" style="position:absolute; left:320px; top:507px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">The building step creates the indexes for both searching</span></div> <div class="txt" style="position:absolute; left:320px; top:518px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and browsing. The MG software is generally used to do the</span></div> <div class="txt" style="position:absolute; left:320px; top:528px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">searching (Witten </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">et al.</span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, 1999), and the </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">mgbuild </span><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">module is</span></div> <div class="txt" style="position:absolute; left:320px; top:539px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">automatically invoked to create each of the indexes that is</span></div> <div class="txt" style="position:absolute; left:320px; top:549px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">required. For example, the Humanity Development Library</span></div> <div class="txt" style="position:absolute; left:320px; top:560px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">has three indexes, one for entire books, one for chapters,</span></div> <div class="txt" style="position:absolute; left:320px; top:571px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and one for section titles. Subdirectories of the </span><span class="p5f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">index</span></div> <div class="txt" style="position:absolute; left:320px; top:581px;"><span class="p5f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">directory are created for each of these indexes.</span></div> </div> </div>

Page 6 <div id="page6"> <div style="position: relative; height: 792px; width: 612px;"><style type="text/css"> .txt { white-space:nowrap; } .p6f0 { font-family:sans-serif; font-weight:normal; font-style:normal; } .p6f1 { font-family:sans-serif; font-weight:normal; font-style:normal; } .p6f2 { font-family:serif; font-weight:normal; font-style:normal; } .p6f3 { font-family:serif; font-weight:normal; font-style:italic; } .p6f4 { font-family:serif; font-weight:normal; font-style:normal; } .p6f5 { font-family:sans-serif; font-weight:bold; font-style:normal; } .p6f6 { font-family:sans-serif; font-weight:normal; font-style:normal; } </style> <img class="background" height="792" id="background6" src="_httpdocimg_/page6.png" style="position:absolute; left:0px; top:0px;" width="612"> <div class="txt" style="position:absolute; left:125px; top:54px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">creator</span></div> <div class="txt" style="position:absolute; left:214px; top:54px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">davidb@cs.waikato.ac.nz</span></div> <div class="txt" style="position:absolute; left:506px; top:54px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">1</span></div> <div class="txt" style="position:absolute; left:125px; top:62px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">maintainer</span></div> <div class="txt" style="position:absolute; left:214px; top:62px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">davidb@cs.waikato.ac.nz</span></div> <div class="txt" style="position:absolute; left:506px; top:62px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">2</span></div> <div class="txt" style="position:absolute; left:125px; top:71px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">public</span></div> <div class="txt" style="position:absolute; left:214px; top:71px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">True</span></div> <div class="txt" style="position:absolute; left:506px; top:71px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">3</span></div> <div class="txt" style="position:absolute; left:506px; top:80px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">4</span></div> <div class="txt" style="position:absolute; left:125px; top:89px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">indexes</span></div> <div class="txt" style="position:absolute; left:214px; top:89px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">document:text</span></div> <div class="txt" style="position:absolute; left:506px; top:89px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">5</span></div> <div class="txt" style="position:absolute; left:125px; top:98px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">defaultindex</span></div> <div class="txt" style="position:absolute; left:214px; top:98px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">document:text</span></div> <div class="txt" style="position:absolute; left:506px; top:98px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">6</span></div> <div class="txt" style="position:absolute; left:125px; top:107px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">plugins</span></div> <div class="txt" style="position:absolute; left:214px; top:107px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">GMLPlug TEXTPlug ArcPlug RecPlug</span></div> <div class="txt" style="position:absolute; left:506px; top:107px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">7</span></div> <div class="txt" style="position:absolute; left:506px; top:116px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">8</span></div> <div class="txt" style="position:absolute; left:125px; top:125px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">classify</span></div> <div class="txt" style="position:absolute; left:214px; top:125px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">AZList metadata=Title</span></div> <div class="txt" style="position:absolute; left:506px; top:125px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">9</span></div> <div class="txt" style="position:absolute; left:506px; top:133px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">10</span></div> <div class="txt" style="position:absolute; left:125px; top:142px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">collectionmeta</span></div> <div class="txt" style="position:absolute; left:214px; top:142px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">collectionname &quot;generic text collection&quot;</span></div> <div class="txt" style="position:absolute; left:506px; top:142px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">11</span></div> <div class="txt" style="position:absolute; left:97px; top:151px;"><span class="p6f2" style="font-size:10px;vertical-align:baseline;color:rgba(0,0,0,1);">(a) </span><span class="p6f2" style="font-size:7px;vertical-align:super;color:rgba(0,0,0,1);">collectionmeta</span></div> <div class="txt" style="position:absolute; left:214px; top:151px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">.document:text &quot;documents&quot;</span></div> <div class="txt" style="position:absolute; left:506px; top:151px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">12</span></div> <div class="txt" style="position:absolute; left:125px; top:177px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">creator</span></div> <div class="txt" style="position:absolute; left:214px; top:177px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">davidb@cs.waikato.ac.nz</span></div> <div class="txt" style="position:absolute; left:506px; top:177px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">1</span></div> <div class="txt" style="position:absolute; left:125px; top:186px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">maintainer</span></div> <div class="txt" style="position:absolute; left:214px; top:186px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">davidb@cs.waikato.ac.nz</span></div> <div class="txt" style="position:absolute; left:506px; top:186px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">2</span></div> <div class="txt" style="position:absolute; left:125px; top:195px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">public</span></div> <div class="txt" style="position:absolute; left:214px; top:195px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">True</span></div> <div class="txt" style="position:absolute; left:506px; top:195px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">3</span></div> <div class="txt" style="position:absolute; left:506px; top:204px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">4</span></div> <div class="txt" style="position:absolute; left:125px; top:213px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">indexes</span></div> <div class="txt" style="position:absolute; left:214px; top:213px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">document:text document:From</span></div> <div class="txt" style="position:absolute; left:506px; top:213px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">5</span></div> <div class="txt" style="position:absolute; left:125px; top:222px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">defaultindex</span></div> <div class="txt" style="position:absolute; left:214px; top:222px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">document:text</span></div> <div class="txt" style="position:absolute; left:506px; top:222px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">6</span></div> <div class="txt" style="position:absolute; left:125px; top:230px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">plugins</span></div> <div class="txt" style="position:absolute; left:214px; top:230px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">GMLPlug EMAILPlug ArcPlug RecPlug</span></div> <div class="txt" style="position:absolute; left:506px; top:230px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">7</span></div> <div class="txt" style="position:absolute; left:506px; top:239px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">8</span></div> <div class="txt" style="position:absolute; left:125px; top:248px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">classify</span></div> <div class="txt" style="position:absolute; left:214px; top:248px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">AZList metadata=Title</span></div> <div class="txt" style="position:absolute; left:506px; top:248px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">9</span></div> <div class="txt" style="position:absolute; left:125px; top:257px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">classify</span></div> <div class="txt" style="position:absolute; left:214px; top:257px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">DateList</span></div> <div class="txt" style="position:absolute; left:506px; top:257px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">10</span></div> <div class="txt" style="position:absolute; left:506px; top:266px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">11</span></div> <div class="txt" style="position:absolute; left:125px; top:275px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">collectionmeta</span></div> <div class="txt" style="position:absolute; left:214px; top:275px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">collectionname &quot;Email messages&quot;</span></div> <div class="txt" style="position:absolute; left:506px; top:275px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">12</span></div> <div class="txt" style="position:absolute; left:125px; top:284px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">collectionmeta</span></div> <div class="txt" style="position:absolute; left:214px; top:284px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">.document:text &quot;documents&quot;</span></div> <div class="txt" style="position:absolute; left:506px; top:284px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">13</span></div> <div class="txt" style="position:absolute; left:125px; top:293px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">collectionmeta</span></div> <div class="txt" style="position:absolute; left:214px; top:293px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">.document:From &quot;email senders&quot;</span></div> <div class="txt" style="position:absolute; left:506px; top:293px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">14</span></div> <div class="txt" style="position:absolute; left:506px; top:301px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">15</span></div> <div class="txt" style="position:absolute; left:125px; top:310px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">format</span></div> <div class="txt" style="position:absolute; left:214px; top:310px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">QueryResults \\</span></div> <div class="txt" style="position:absolute; left:506px; top:310px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">16</span></div> <div class="txt" style="position:absolute; left:97px; top:320px;"><span class="p6f2" style="font-size:10px;vertical-align:baseline;color:rgba(0,0,0,1);">(b)</span></div> <div class="txt" style="position:absolute; left:214px; top:319px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">&lt;td&gt;[link][icon][/link]&lt;/td&gt;&lt;td&gt;[Title]&lt;/td&gt;&lt;td&gt;[Author]&lt;/td&gt;</span></div> <div class="txt" style="position:absolute; left:506px; top:319px;"><span class="p6f2" style="font-size:7px;vertical-align:baseline;color:rgba(0,0,0,1);">17</span></div> <div class="txt" style="position:absolute; left:178px; top:338px;"><span class="p6f5" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">Figure 5: Collection configuration files (a) generic, (b) for an email collection</span></div> <div class="txt" style="position:absolute; left:63px; top:365px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">MG also compresses the text of the collection; and the</span></div> <div class="txt" style="position:absolute; left:63px; top:375px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">image files are linked into the </span><span class="p6f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">index </span><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">subdirectory. Now</span></div> <div class="txt" style="position:absolute; left:63px; top:386px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">none of the material in the </span><span class="p6f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">import </span><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and </span><span class="p6f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">archives </span><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">directories</span></div> <div class="txt" style="position:absolute; left:63px; top:397px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">is needed to run the collection and can be removed from</span></div> <div class="txt" style="position:absolute; left:63px; top:407px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the file system (though they would be needed if the</span></div> <div class="txt" style="position:absolute; left:63px; top:418px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">collection were rebuilt).</span></div> <div class="txt" style="position:absolute; left:63px; top:434px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Associated with each collection is a database stored in</span></div> <div class="txt" style="position:absolute; left:63px; top:445px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">GDBM (Gnu database manager) format. This contains an</span></div> <div class="txt" style="position:absolute; left:63px; top:455px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">entry for each document, giving its OID, its internal MG</span></div> <div class="txt" style="position:absolute; left:63px; top:466px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">document number, and metadata such as title. Information</span></div> <div class="txt" style="position:absolute; left:63px; top:476px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">for each of the browsing indexes, which appear as buttons</span></div> <div class="txt" style="position:absolute; left:63px; top:487px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">on the Greenstone search/browse bar, is also extracted</span></div> <div class="txt" style="position:absolute; left:63px; top:497px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">during the building process and stored in the database. A</span></div> <div class="txt" style="position:absolute; left:63px; top:508px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">“classifier” program is required for each browsing index to</span></div> <div class="txt" style="position:absolute; left:63px; top:519px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">extract the appropriate information from GML documents.</span></div> <div class="txt" style="position:absolute; left:63px; top:529px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Like plugins, classifiers are written on an </span><span class="p6f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">ad hoc </span><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">basis for</span></div> <div class="txt" style="position:absolute; left:63px; top:540px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the particular information required, and where possible</span></div> <div class="txt" style="position:absolute; left:63px; top:550px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">reused from one collection to another.</span></div> <div class="txt" style="position:absolute; left:63px; top:567px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">The building program creates the indexes based on</span></div> <div class="txt" style="position:absolute; left:63px; top:577px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">whatever appears in the </span><span class="p6f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">archives </span><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">directory. The first plugin</span></div> <div class="txt" style="position:absolute; left:63px; top:588px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">specified by all collections is one that processes GML</span></div> <div class="txt" style="position:absolute; left:63px; top:598px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">files, and so if </span><span class="p6f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">archives </span><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">contains imported files they will be</span></div> <div class="txt" style="position:absolute; left:63px; top:609px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">processed correctly. If it contains material in the original</span></div> <div class="txt" style="position:absolute; left:63px; top:619px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">format, that will be converted using the appropriate plugin.</span></div> <div class="txt" style="position:absolute; left:63px; top:630px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Thus the import process is optional.</span></div> <div class="txt" style="position:absolute; left:63px; top:646px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">GML is designed to be fast and easy to parse, an important</span></div> <div class="txt" style="position:absolute; left:63px; top:657px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">requirement when millions of documents are to be</span></div> <div class="txt" style="position:absolute; left:63px; top:667px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">processed. Something as simple as requiring tags to be</span></div> <div class="txt" style="position:absolute; left:63px; top:678px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">lower-case, for example, yields a substantial speed-up. In</span></div> <div class="txt" style="position:absolute; left:320px; top:365px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">certain circumstances, however, it might be preferable to</span></div> <div class="txt" style="position:absolute; left:320px; top:375px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">use a standardized format such as XML. This is</span></div> <div class="txt" style="position:absolute; left:320px; top:385px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">straightforward to implement</span><span class="p6f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);"></span><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">just write an XML</span></div> <div class="txt" style="position:absolute; left:320px; top:396px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">plugin</span><span class="p6f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);"></span><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">although we have not done so ourselves. Given</span></div> <div class="txt" style="position:absolute; left:320px; top:407px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the transitory nature of the imported data, to date, we have</span></div> <div class="txt" style="position:absolute; left:320px; top:418px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">found GML a satisfactory and beneficial format.</span></div> <div class="txt" style="position:absolute; left:320px; top:440px;"><span class="p6f5" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">CREATING NEW COLLECTIONS</span></div> <div class="txt" style="position:absolute; left:320px; top:456px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Building new collections from scratch is only slightly</span></div> <div class="txt" style="position:absolute; left:320px; top:466px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">different from updating an existing collection. The key</span></div> <div class="txt" style="position:absolute; left:320px; top:477px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">new requirement is creating a collection configuration file,</span></div> <div class="txt" style="position:absolute; left:320px; top:488px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and a software utility is provided to help. Two pieces of</span></div> <div class="txt" style="position:absolute; left:320px; top:498px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">information are required for this: the name of the directory</span></div> <div class="txt" style="position:absolute; left:320px; top:509px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">that the collection will use (into which the source data and</span></div> <div class="txt" style="position:absolute; left:320px; top:519px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">other files will eventually be placed), and a contact e-mail</span></div> <div class="txt" style="position:absolute; left:320px; top:530px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">address for use if any problems are encountered by the</span></div> <div class="txt" style="position:absolute; left:320px; top:540px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">software once the collection is up and running. The utility</span></div> <div class="txt" style="position:absolute; left:320px; top:551px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">creates files and directories within the newly-named</span></div> <div class="txt" style="position:absolute; left:320px; top:561px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">directory to support a generic collection of plain text</span></div> <div class="txt" style="position:absolute; left:320px; top:572px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">documents. With suitable data placed in the </span><span class="p6f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">import</span></div> <div class="txt" style="position:absolute; left:320px; top:583px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">directory, building the collection at this point will yield a</span></div> <div class="txt" style="position:absolute; left:320px; top:593px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">document-level searchable index of all the text and a</span></div> <div class="txt" style="position:absolute; left:320px; top:604px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">browsable list of “titles” (defined in this case to be the</span></div> <div class="txt" style="position:absolute; left:320px; top:614px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">document filenames).</span></div> <div class="txt" style="position:absolute; left:320px; top:631px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">To enhance the functionality and presentation— something</span></div> <div class="txt" style="position:absolute; left:320px; top:641px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">anything but the most trivial collection will require—the</span></div> <div class="txt" style="position:absolute; left:320px; top:652px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">configuration file must be edited. For a collection sourced</span></div> <div class="txt" style="position:absolute; left:320px; top:662px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">from documents in an already supported data format,</span></div> <div class="txt" style="position:absolute; left:320px; top:673px;"><span class="p6f2" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">presented in a similar fashion to an existing collection, the</span></div> </div> </div>

Page 7 <div id="page7"> <div style="position: relative; height: 792px; width: 612px;"><style type="text/css"> .txt { white-space:nowrap; } .p7f0 { font-family:sans-serif; font-weight:normal; font-style:normal; } .p7f1 { font-family:serif; font-weight:normal; font-style:normal; } .p7f2 { font-family:sans-serif; font-weight:normal; font-style:normal; } .p7f3 { font-family:sans-serif; font-weight:bold; font-style:normal; } .p7f4 { font-family:serif; font-weight:normal; font-style:italic; } .p7f5 { font-family:sans-serif; font-weight:normal; font-style:normal; } </style> <img class="background" height="792" id="background7" src="_httpdocimg_/page7.png" style="position:absolute; left:0px; top:0px;" width="612"> <div class="txt" style="position:absolute; left:99px; top:286px;"><span class="p7f3" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">Figure 6: Searching bookmarked Web pages</span></div> <div class="txt" style="position:absolute; left:63px; top:315px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">amount of editing is minimal. Importing new data formats</span></div> <div class="txt" style="position:absolute; left:63px; top:325px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and browsing metadata in ways not currently supported are</span></div> <div class="txt" style="position:absolute; left:63px; top:336px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">more complex activities that require programming skills.</span></div> <div class="txt" style="position:absolute; left:63px; top:366px;"><span class="p7f3" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">Modifying the configuration file</span></div> <div class="txt" style="position:absolute; left:63px; top:381px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Figure 5b shows simple alterations to the generic</span></div> <div class="txt" style="position:absolute; left:63px; top:392px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">configuration file in Figure 5a that was generated by the</span></div> <div class="txt" style="position:absolute; left:63px; top:403px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">new-collection utility. </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">TEXTPlug </span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">is replaced with</span></div> <div class="txt" style="position:absolute; left:63px; top:413px;"><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">EMAILPlug </span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">(line 7) which reads email files and extracts</span></div> <div class="txt" style="position:absolute; left:63px; top:424px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">metadata (</span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">From</span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">To</span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Date</span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Subject</span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">) from them. A classifier</span></div> <div class="txt" style="position:absolute; left:63px; top:434px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">for dates is added (line 10) to make the collection</span></div> <div class="txt" style="position:absolute; left:63px; top:445px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">browsable chronologically. The default presentation of</span></div> <div class="txt" style="position:absolute; left:63px; top:455px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">search results is overridden (line 17) to display both the</span></div> <div class="txt" style="position:absolute; left:63px; top:466px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">title of the message (i.e. Dublin Core </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Title</span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">) and its sender</span></div> <div class="txt" style="position:absolute; left:63px; top:477px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">(i.e. Dublin Core </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Author</span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">). Elements in square brackets,</span></div> <div class="txt" style="position:absolute; left:63px; top:487px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">such as </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">[Title]</span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, are replaced by the metadata associated</span></div> <div class="txt" style="position:absolute; left:63px; top:498px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">with a particular document. The built-in term </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">[icon]</span></div> <div class="txt" style="position:absolute; left:63px; top:508px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">produces a suitable image that represents the document</span></div> <div class="txt" style="position:absolute; left:63px; top:519px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">(such as a book icon or page icon), and the </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">[link]…[/link]</span></div> <div class="txt" style="position:absolute; left:63px; top:529px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">construct forms a hyperlink to the complete document.</span></div> <div class="txt" style="position:absolute; left:63px; top:540px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Anything else in the format statement, which in this case is</span></div> <div class="txt" style="position:absolute; left:63px; top:550px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">solely table-cell tags in HTML, is passed through to the</span></div> <div class="txt" style="position:absolute; left:63px; top:561px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">page being displayed.</span></div> <div class="txt" style="position:absolute; left:63px; top:577px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">As this example shows, creating a new collection that stays</span></div> <div class="txt" style="position:absolute; left:63px; top:588px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">within the bounds of the library’s established capabilities</span></div> <div class="txt" style="position:absolute; left:63px; top:598px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">falls within the capability of many computer users—for</span></div> <div class="txt" style="position:absolute; left:63px; top:609px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">instance, computer-trained librarians. Extending</span></div> <div class="txt" style="position:absolute; left:63px; top:620px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Greenstone to handle new document formats and browse</span></div> <div class="txt" style="position:absolute; left:63px; top:630px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">metadata in new ways is more challenging.</span></div> <div class="txt" style="position:absolute; left:63px; top:660px;"><span class="p7f3" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">Writing new plugins and classifiers</span></div> <div class="txt" style="position:absolute; left:63px; top:676px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Extensibility is obtained through plugins and classifiers.</span></div> <div class="txt" style="position:absolute; left:320px; top:53px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">These are modules of code that can be slotted into the</span></div> <div class="txt" style="position:absolute; left:320px; top:64px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">system to enhance its capabilities. Plugins parse</span></div> <div class="txt" style="position:absolute; left:320px; top:74px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">documents, extracting the text and metadata to be indexed.</span></div> <div class="txt" style="position:absolute; left:320px; top:85px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Classifiers control how metadata is brought together to</span></div> <div class="txt" style="position:absolute; left:320px; top:95px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">form browsable data structures. Both are specified in an</span></div> <div class="txt" style="position:absolute; left:320px; top:106px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">object-oriented framework using inheritance to minimize</span></div> <div class="txt" style="position:absolute; left:320px; top:117px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the amount of code written.</span></div> <div class="txt" style="position:absolute; left:320px; top:133px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">A plugin must specify three things: what file formats it can</span></div> <div class="txt" style="position:absolute; left:320px; top:143px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">handle, how they should be parsed, and whether the plugin</span></div> <div class="txt" style="position:absolute; left:320px; top:154px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">is recursive. File formats are normally determined using</span></div> <div class="txt" style="position:absolute; left:320px; top:165px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">regular expression matching on the filename. For example,</span></div> <div class="txt" style="position:absolute; left:320px; top:175px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the HTML plugin accepts all files that end in </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">.htm</span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, .</span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">html</span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">,</span></div> <div class="txt" style="position:absolute; left:320px; top:186px;"><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">.HTM</span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, or </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">.HTML</span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">. (It is quite possible, however, to write</span></div> <div class="txt" style="position:absolute; left:320px; top:196px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">plugins that “look inside” the file as well.) For other files,</span></div> <div class="txt" style="position:absolute; left:320px; top:207px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the plugin returns </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">undefined </span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and the file is passed to the</span></div> <div class="txt" style="position:absolute; left:320px; top:217px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">next plugin in the collection’s configuration file (e.g.</span></div> <div class="txt" style="position:absolute; left:320px; top:228px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Figure 5 line 7). If it can, the plugin parses the file and</span></div> <div class="txt" style="position:absolute; left:320px; top:238px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">returns the number of documents processed. This involves</span></div> <div class="txt" style="position:absolute; left:320px; top:249px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">extracting text and metadata and adding it to the library’s</span></div> <div class="txt" style="position:absolute; left:320px; top:260px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">content through calls to </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">add text </span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">add metadata</span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">.</span></div> <div class="txt" style="position:absolute; left:320px; top:276px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Some plugins (“recursive” ones) add extra files into the</span></div> <div class="txt" style="position:absolute; left:320px; top:286px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">stream of data processed during the building phase by</span></div> <div class="txt" style="position:absolute; left:320px; top:297px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">artificially reactivating the list of plugins. This is how</span></div> <div class="txt" style="position:absolute; left:320px; top:308px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">directory hierarchies are traversed.</span></div> <div class="txt" style="position:absolute; left:320px; top:324px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Plugins are small modules of code that are easy to write.</span></div> <div class="txt" style="position:absolute; left:320px; top:334px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">We monitored the time it took to develop a new one that</span></div> <div class="txt" style="position:absolute; left:320px; top:345px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">was different to any we had produced so far. We chose to</span></div> <div class="txt" style="position:absolute; left:320px; top:356px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">make as an example a collection of HTML bookmark files,</span></div> <div class="txt" style="position:absolute; left:320px; top:366px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the motivation being to produce a convenient way of</span></div> <div class="txt" style="position:absolute; left:320px; top:377px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">searching and browsing one’s bookmarked Web pages.</span></div> <div class="txt" style="position:absolute; left:320px; top:387px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Figure 6 shows a user searching for bookmarked pages</span></div> <div class="txt" style="position:absolute; left:320px; top:398px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">about </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">music</span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">. The new plugin took under an hour to write,</span></div> <div class="txt" style="position:absolute; left:320px; top:408px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and was 160 lines long (ignoring blank lines and</span></div> <div class="txt" style="position:absolute; left:320px; top:419px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">comments)—about the average length of existing plugins.</span></div> <div class="txt" style="position:absolute; left:320px; top:435px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Classifiers are more general than plugins because they</span></div> <div class="txt" style="position:absolute; left:320px; top:446px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">work on GML-format data. For example, any plugin that</span></div> <div class="txt" style="position:absolute; left:320px; top:456px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">generates date metadata in accordance with the Dublin</span></div> <div class="txt" style="position:absolute; left:320px; top:467px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">core can request the collection to be browsable</span></div> <div class="txt" style="position:absolute; left:320px; top:477px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">chronologically by specifying the </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">DateList </span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">classifier in the</span></div> <div class="txt" style="position:absolute; left:320px; top:488px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">collection’s configuration file (Figure 7). Classifiers are</span></div> <div class="txt" style="position:absolute; left:320px; top:499px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">more elaborate than most plugins, but new ones are seldom</span></div> <div class="txt" style="position:absolute; left:320px; top:509px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">required. The average length of existing classifiers is 230</span></div> <div class="txt" style="position:absolute; left:320px; top:520px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">lines.</span></div> <div class="txt" style="position:absolute; left:320px; top:536px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Classifiers must specify three things: an initialization</span></div> <div class="txt" style="position:absolute; left:320px; top:547px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">routine, how individual documents are classified, and the</span></div> <div class="txt" style="position:absolute; left:320px; top:557px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">final browsable data structure. Initialization takes care of</span></div> <div class="txt" style="position:absolute; left:320px; top:568px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">any options specified in the configuration file (such as</span></div> <div class="txt" style="position:absolute; left:320px; top:578px;"><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">metadata=Title </span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">on line 9 of Figure 5b). Classifying</span></div> <div class="txt" style="position:absolute; left:320px; top:589px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">individual documents is an iterative process: for each one,</span></div> <div class="txt" style="position:absolute; left:320px; top:599px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">a call to </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">document-classify </span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">is made. On presentation of the</span></div> <div class="txt" style="position:absolute; left:320px; top:610px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">document’s OID, the necessary metadata is located and</span></div> <div class="txt" style="position:absolute; left:320px; top:621px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">used to control where the document is added to the</span></div> <div class="txt" style="position:absolute; left:320px; top:631px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">browsable data structure being constructed.</span></div> <div class="txt" style="position:absolute; left:320px; top:647px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Once all documents have been added, a request is made for</span></div> <div class="txt" style="position:absolute; left:320px; top:658px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the completed data structure. Some classifiers return the</span></div> <div class="txt" style="position:absolute; left:320px; top:669px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">data structure directly; others transform the data structure</span></div> <div class="txt" style="position:absolute; left:320px; top:679px;"><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">before it is returned. For example, the </span><span class="p7f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">AZList </span><span class="p7f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">classifier</span></div> </div> </div>

Page 8 <div id="page8"> <div style="position: relative; height: 792px; width: 612px;"><style type="text/css"> .txt { white-space:nowrap; } .p8f0 { font-family:sans-serif; font-weight:normal; font-style:normal; } .p8f1 { font-family:serif; font-weight:normal; font-style:normal; } .p8f2 { font-family:sans-serif; font-weight:bold; font-style:normal; } .p8f3 { font-family:serif; font-weight:normal; font-style:italic; } .p8f4 { font-family:serif; font-weight:normal; font-style:normal; } .p8f5 { font-family:sans-serif; font-weight:normal; font-style:normal; } </style> <img class="background" height="792" id="background8" src="_httpdocimg_/page8.png" style="position:absolute; left:0px; top:0px;" width="612"> <div class="txt" style="position:absolute; left:88px; top:287px;"><span class="p8f2" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">Figure 7: Browsing a newspaper collection by date</span></div> <div class="txt" style="position:absolute; left:63px; top:315px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">divides the alphabetically sorted list of metadata into</span></div> <div class="txt" style="position:absolute; left:63px; top:325px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">separate pages of about the same size and returns the</span></div> <div class="txt" style="position:absolute; left:63px; top:336px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">alphabetic ranges for each one (Figure 4).</span></div> <div class="txt" style="position:absolute; left:63px; top:358px;"><span class="p8f2" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">OVERVIEW OF RELATED WORK</span></div> <div class="txt" style="position:absolute; left:63px; top:374px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Two projects that provide substantial open source digital</span></div> <div class="txt" style="position:absolute; left:63px; top:385px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">library software are Dienst (Lagoze and Fielding, 1998)</span></div> <div class="txt" style="position:absolute; left:63px; top:396px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and Harvest (Bowman </span><span class="p8f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">et al.</span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, 1994). The origins of Dienst</span></div> <div class="txt" style="position:absolute; left:63px; top:407px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">(</span><span class="p8f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">www.cs.cornell.edu/cdlrg</span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">) stretch back to 1992. The term</span></div> <div class="txt" style="position:absolute; left:63px; top:418px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">has come to represent three entities: a conceptual</span></div> <div class="txt" style="position:absolute; left:63px; top:429px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">architecture for distributed digital libraries; an open</span></div> <div class="txt" style="position:absolute; left:63px; top:441px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">protocol for service communication; and a software</span></div> <div class="txt" style="position:absolute; left:63px; top:452px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">system that implements the protocol. To date, five sample</span></div> <div class="txt" style="position:absolute; left:63px; top:463px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">digital libraries have been built using this technology.</span></div> <div class="txt" style="position:absolute; left:63px; top:474px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">They manifest themselves in two forms: technical reports</span></div> <div class="txt" style="position:absolute; left:63px; top:485px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and primary source documents.</span></div> <div class="txt" style="position:absolute; left:63px; top:502px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Best known is NCSTRL, the Networked Computer</span></div> <div class="txt" style="position:absolute; left:63px; top:513px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Science Technical Reference Library project</span></div> <div class="txt" style="position:absolute; left:63px; top:524px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">(</span><span class="p8f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">www.ncstrl.org</span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">). This collection facilitates searching by</span></div> <div class="txt" style="position:absolute; left:63px; top:535px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">title, author and abstract, and browsing by year and author,</span></div> <div class="txt" style="position:absolute; left:63px; top:546px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">across a distributed network of document repositories.</span></div> <div class="txt" style="position:absolute; left:63px; top:557px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Documents can (where supported) be delivered in various</span></div> <div class="txt" style="position:absolute; left:63px; top:568px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">formats such as PostScript, a thumbnail overview of the</span></div> <div class="txt" style="position:absolute; left:63px; top:579px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">pages, and a GIF image of a particular page.</span></div> <div class="txt" style="position:absolute; left:63px; top:596px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">The </span><span class="p8f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Making of America </span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">resource is an example of a</span></div> <div class="txt" style="position:absolute; left:63px; top:606px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">collection based around primary sources</span><span class="p8f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);"></span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">in this case</span></div> <div class="txt" style="position:absolute; left:63px; top:618px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">American social history, 1830</span><span class="p8f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">−</span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">1900. It has a different</span></div> <div class="txt" style="position:absolute; left:63px; top:630px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">“look and feel” to NCSTRL, being strongly oriented</span></div> <div class="txt" style="position:absolute; left:63px; top:641px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">toward browsing rather than searching. A user navigates</span></div> <div class="txt" style="position:absolute; left:63px; top:652px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">their way through a hierarchical structure of hyperlinks to</span></div> <div class="txt" style="position:absolute; left:63px; top:664px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">reach a book of interest. The book itself is a series of</span></div> <div class="txt" style="position:absolute; left:63px; top:675px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">scanned images: delivery options include going directly to</span></div> <div class="txt" style="position:absolute; left:320px; top:53px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">a page number, next and previous page buttons, and</span></div> <div class="txt" style="position:absolute; left:320px; top:64px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">displaying a particular page at different resolutions. A text</span></div> <div class="txt" style="position:absolute; left:320px; top:75px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">version of the page is also available upon which a</span></div> <div class="txt" style="position:absolute; left:320px; top:87px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">searching option is also provided.</span></div> <div class="txt" style="position:absolute; left:320px; top:103px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Started in 1994, Harvest is also a long-running research</span></div> <div class="txt" style="position:absolute; left:320px; top:115px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">project. It provides an efficient means of gathering source</span></div> <div class="txt" style="position:absolute; left:320px; top:126px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">data from the Internet and distributing indexing</span></div> <div class="txt" style="position:absolute; left:320px; top:137px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">information over the Internet. This is accomplished</span></div> <div class="txt" style="position:absolute; left:320px; top:148px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">through five components: </span><span class="p8f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">gatherer</span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, </span><span class="p8f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">broker</span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, </span><span class="p8f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">indexer</span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">,</span></div> <div class="txt" style="position:absolute; left:320px; top:159px;"><span class="p8f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">replicator </span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">and </span><span class="p8f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">cache</span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">. The first three are central to creating,</span></div> <div class="txt" style="position:absolute; left:320px; top:170px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">updating and searching a collection; the last two help to</span></div> <div class="txt" style="position:absolute; left:320px; top:181px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">improve performance over the Internet through transparent</span></div> <div class="txt" style="position:absolute; left:320px; top:192px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">mirroring and caching techniques.</span></div> <div class="txt" style="position:absolute; left:320px; top:209px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">The system is configurable and customizable. While</span></div> <div class="txt" style="position:absolute; left:320px; top:220px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">searching is most commonly implemented using Glimpse</span></div> <div class="txt" style="position:absolute; left:320px; top:231px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">(</span><span class="p8f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">glimpse.cs.arizona.edu</span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">), in principle any search engine</span></div> <div class="txt" style="position:absolute; left:320px; top:242px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">that supports incremental updates and Boolean</span></div> <div class="txt" style="position:absolute; left:320px; top:253px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">combinations of attribute-based queries can be used. It is</span></div> <div class="txt" style="position:absolute; left:320px; top:264px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">possible to control what type of documents are gathered</span></div> <div class="txt" style="position:absolute; left:320px; top:275px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">during creation and updating, and how the query interface</span></div> <div class="txt" style="position:absolute; left:320px; top:286px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">looks and is laid out.</span></div> <div class="txt" style="position:absolute; left:320px; top:303px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Sample collections cited by the developers include 21,000</span></div> <div class="txt" style="position:absolute; left:320px; top:315px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">computer science technical reports and 7,000 home pages.</span></div> <div class="txt" style="position:absolute; left:320px; top:326px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Other examples include a sizable collection of agriculture-</span></div> <div class="txt" style="position:absolute; left:320px; top:337px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">related electronic journals and magazines called “tomato-</span></div> <div class="txt" style="position:absolute; left:320px; top:348px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">juice” (accessed through </span><span class="p8f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">hegel.lib.ncsu.edu</span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">) and a full-text</span></div> <div class="txt" style="position:absolute; left:320px; top:359px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">index of library-related electronic serials</span></div> <div class="txt" style="position:absolute; left:320px; top:370px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">(</span><span class="p8f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">sunsite.berkeley.edu/IndexMorganagus</span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">). Harvest is also</span></div> <div class="txt" style="position:absolute; left:320px; top:381px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">often used to index Web sites (for example</span></div> <div class="txt" style="position:absolute; left:320px; top:392px;"><span class="p8f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">www.middlebury.edu</span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">).</span></div> <div class="txt" style="position:absolute; left:320px; top:409px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Comparing Greenstone with Dienst and Harvest, there are</span></div> <div class="txt" style="position:absolute; left:320px; top:420px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">both similarities and differences. All provide substantial</span></div> <div class="txt" style="position:absolute; left:320px; top:431px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">digital library systems, hence common themes recur, but</span></div> <div class="txt" style="position:absolute; left:320px; top:442px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">they are driven by projects with different aims. Harvest,</span></div> <div class="txt" style="position:absolute; left:320px; top:453px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">for instance, was not conceived as a digital library project</span></div> <div class="txt" style="position:absolute; left:320px; top:464px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">at all, but by virtue of its selective document gathering</span></div> <div class="txt" style="position:absolute; left:320px; top:475px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">process it can be classed (and is used) as one. While it</span></div> <div class="txt" style="position:absolute; left:320px; top:487px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">provides sophisticated search options, it lacks the</span></div> <div class="txt" style="position:absolute; left:320px; top:498px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">complementary service of browsing. Furthermore it adds</span></div> <div class="txt" style="position:absolute; left:320px; top:509px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">no structure or order to the documents collected, relying</span></div> <div class="txt" style="position:absolute; left:320px; top:520px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">on whatever structures are present in the site that they</span></div> <div class="txt" style="position:absolute; left:320px; top:531px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">were gathered from. A proven strength of the design is its</span></div> <div class="txt" style="position:absolute; left:320px; top:541px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">flexibility through configuration and customization</span><span class="p8f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);"></span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">an</span></div> <div class="txt" style="position:absolute; left:320px; top:554px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">element also present in Greenstone.</span></div> <div class="txt" style="position:absolute; left:320px; top:570px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Dienst</span><span class="p8f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);"></span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">best exemplified through the NCSTRL</span></div> <div class="txt" style="position:absolute; left:320px; top:581px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">work</span><span class="p8f4" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);"></span><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">supports searching and browsing, like Greenstone.</span></div> <div class="txt" style="position:absolute; left:320px; top:593px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Both use open protocols. Differences include a high</span></div> <div class="txt" style="position:absolute; left:320px; top:604px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">reliance in Dienst on user-supplied information when a</span></div> <div class="txt" style="position:absolute; left:320px; top:616px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">document is added, and a smaller range of document types</span></div> <div class="txt" style="position:absolute; left:320px; top:627px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">supported—although Dienst does include a document</span></div> <div class="txt" style="position:absolute; left:320px; top:638px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">model that should, over time, allow this to expand with</span></div> <div class="txt" style="position:absolute; left:320px; top:649px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">relative ease.</span></div> <div class="txt" style="position:absolute; left:320px; top:666px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">There are also commercial systems that provide similar</span></div> <div class="txt" style="position:absolute; left:320px; top:677px;"><span class="p8f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">digital library services to those described. However, since</span></div> </div> </div>

Page 9 <div id="page9"> <div style="position: relative; height: 792px; width: 612px;"><style type="text/css"> .txt { white-space:nowrap; } .p9f0 { font-family:sans-serif; font-weight:normal; font-style:normal; } .p9f1 { font-family:serif; font-weight:normal; font-style:normal; } .p9f2 { font-family:sans-serif; font-weight:bold; font-style:normal; } .p9f3 { font-family:serif; font-weight:normal; font-style:normal; } .p9f4 { font-family:sans-serif; font-weight:normal; font-style:normal; } .p9f5 { font-family:serif; font-weight:normal; font-style:italic; } .p9f6 { font-family:serif; font-weight:normal; font-style:normal; } .p9f7 { font-family:serif; font-weight:normal; font-style:italic; } .p9f8 { font-family:sans-serif; font-weight:normal; font-style:normal; } </style> <img class="background" height="792" id="background9" src="_httpdocimg_/page9.png" style="position:absolute; left:0px; top:0px;" width="612"> <div class="txt" style="position:absolute; left:63px; top:53px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">corporate culture instills proprietary attitudes there is little</span></div> <div class="txt" style="position:absolute; left:63px; top:64px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">opportunity for advancement through a shared</span></div> <div class="txt" style="position:absolute; left:63px; top:75px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">collaborative effort. Consequently they are not reviewed</span></div> <div class="txt" style="position:absolute; left:63px; top:87px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">here.</span></div> <div class="txt" style="position:absolute; left:63px; top:109px;"><span class="p9f2" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">CONCLUSIONS</span></div> <div class="txt" style="position:absolute; left:63px; top:125px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Greenstone is a comprehensive software system for</span></div> <div class="txt" style="position:absolute; left:63px; top:136px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">creating digital library collections. It builds data structures</span></div> <div class="txt" style="position:absolute; left:63px; top:146px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">for searching and browsing from the material provided,</span></div> <div class="txt" style="position:absolute; left:63px; top:157px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">rather than relying on any hand-crafting. The process is</span></div> <div class="txt" style="position:absolute; left:63px; top:167px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">controlled by a configuration file, and once a collection</span></div> <div class="txt" style="position:absolute; left:63px; top:178px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">exists new material can be added completely</span></div> <div class="txt" style="position:absolute; left:63px; top:189px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">automatically. Browsing is based on Dublin Core</span></div> <div class="txt" style="position:absolute; left:63px; top:199px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">metadata.</span></div> <div class="txt" style="position:absolute; left:63px; top:215px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">New collections can be developed easily, particularly if</span></div> <div class="txt" style="position:absolute; left:63px; top:226px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">they resemble existing ones. Extensibility is achieved</span></div> <div class="txt" style="position:absolute; left:63px; top:237px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">through software “plugins” that can be written to</span></div> <div class="txt" style="position:absolute; left:63px; top:247px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">accommodate documents, and metadata, in different</span></div> <div class="txt" style="position:absolute; left:63px; top:258px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">formats. Standard plugins exist for many document types;</span></div> <div class="txt" style="position:absolute; left:63px; top:268px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">new ones are easily written. Browsing is controlled by</span></div> <div class="txt" style="position:absolute; left:63px; top:279px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">“classifiers” that process metadata into browsing structures</span></div> <div class="txt" style="position:absolute; left:63px; top:289px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">(by date, alphabetical, hierarchical, etc).</span></div> <div class="txt" style="position:absolute; left:63px; top:306px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">However, the most powerful support for extensibility is</span></div> <div class="txt" style="position:absolute; left:63px; top:316px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">achieved not by technical means but by making the source</span></div> <div class="txt" style="position:absolute; left:63px; top:327px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">code freely available under the Gnu public license. Only</span></div> <div class="txt" style="position:absolute; left:63px; top:337px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">through an international cooperative effort will digital</span></div> <div class="txt" style="position:absolute; left:63px; top:348px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">library software become sufficiently comprehensive to</span></div> <div class="txt" style="position:absolute; left:63px; top:358px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">meet the world’s needs with the richness and flexibility</span></div> <div class="txt" style="position:absolute; left:63px; top:369px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">that users deserve.</span></div> <div class="txt" style="position:absolute; left:63px; top:391px;"><span class="p9f2" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">ACKNOWLEDGMENTS</span></div> <div class="txt" style="position:absolute; left:63px; top:407px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">We gratefully acknowledge all those who have worked on</span></div> <div class="txt" style="position:absolute; left:63px; top:418px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">the Greenstone software, and all members of the New</span></div> <div class="txt" style="position:absolute; left:63px; top:429px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Zealand Digital Library project for their enthusiasm and</span></div> <div class="txt" style="position:absolute; left:63px; top:439px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">ideas.</span></div> <div class="txt" style="position:absolute; left:320px; top:54px;"><span class="p9f2" style="font-size:8px;vertical-align:baseline;color:rgba(0,0,0,1);">REFERENCES</span></div> <div class="txt" style="position:absolute; left:320px; top:67px;"><span class="p9f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">1. Akscyn, R.M. and Witten, I.H. (1998) “Report on First</span></div> <div class="txt" style="position:absolute; left:333px; top:78px;"><span class="p9f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Summit on International Cooperation on Digital</span></div> <div class="txt" style="position:absolute; left:333px; top:89px;"><span class="p9f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Libraries.” ks.com/idla-wp-oct98.</span></div> <div class="txt" style="position:absolute; left:320px; top:103px;"><span class="p9f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">2. Bowman, C.M., Danzig, P.B., Manber, U., and</span></div> <div class="txt" style="position:absolute; left:333px; top:114px;"><span class="p9f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Schwartz, M.F. “Scalable Internet resource discovery:</span></div> <div class="txt" style="position:absolute; left:333px; top:125px;"><span class="p9f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Research problems and approaches” </span><span class="p9f5" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Communications</span></div> <div class="txt" style="position:absolute; left:333px; top:135px;"><span class="p9f5" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">of the ACM, </span><span class="p9f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Vol. 37, No. 8, pp. 98</span><span class="p9f6" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">−</span><span class="p9f3" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">107, 1994.</span></div> <div class="txt" style="position:absolute; left:320px; top:150px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">3. Fox, E. (1998) “Digital library definitions.”</span></div> <div class="txt" style="position:absolute; left:333px; top:161px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">ei.cs.vt.edu/~fox/dlib/def.html.</span></div> <div class="txt" style="position:absolute; left:320px; top:175px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">4. Humanity Libraries (1998) </span><span class="p9f7" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Humanity Development</span></div> <div class="txt" style="position:absolute; left:333px; top:187px;"><span class="p9f7" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Library</span><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">. CD-ROM produced by the Global Help</span></div> <div class="txt" style="position:absolute; left:333px; top:198px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Project, Antwerp, Belgium.</span></div> <div class="txt" style="position:absolute; left:320px; top:212px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">5. Lagoze, C. and Fielding, D “Defining Collections in</span></div> <div class="txt" style="position:absolute; left:333px; top:223px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Distributed Digital Libraries” </span><span class="p9f7" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">D-Lib Magazine</span><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, Nov.</span></div> <div class="txt" style="position:absolute; left:333px; top:234px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">1998.</span></div> <div class="txt" style="position:absolute; left:320px; top:248px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">6. PAHO (1999) </span><span class="p9f7" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Virtual Disaster Library</span><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">. CD-ROM</span></div> <div class="txt" style="position:absolute; left:333px; top:259px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">produced by the Pan-American Health Organization,</span></div> <div class="txt" style="position:absolute; left:333px; top:270px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Washington DC, USA.</span></div> <div class="txt" style="position:absolute; left:320px; top:284px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">7. McNab, R.J., Witten, I.H. and Boddie, S.J. (1998) “A</span></div> <div class="txt" style="position:absolute; left:333px; top:295px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">distributed digital library architecture incorporating</span></div> <div class="txt" style="position:absolute; left:333px; top:306px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">different index styles.” </span><span class="p9f7" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Proc IEEE Advances in Digital</span></div> <div class="txt" style="position:absolute; left:333px; top:317px;"><span class="p9f7" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Libraries</span><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, Santa Barbara, CA, pp. 36–45.</span></div> <div class="txt" style="position:absolute; left:320px; top:331px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">8. Nevill-Manning, C.G., Reed, T., and Witten, I.H.</span></div> <div class="txt" style="position:absolute; left:333px; top:342px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">(1998) “Extracting text from PostScript”</span></div> <div class="txt" style="position:absolute; left:333px; top:353px;"><span class="p9f7" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Software—Practice and Experience</span><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, Vol. 28, No. 5, pp.</span></div> <div class="txt" style="position:absolute; left:333px; top:364px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">481–491; April.</span></div> <div class="txt" style="position:absolute; left:320px; top:378px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">9. UNESCO (1999) </span><span class="p9f7" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">SAHEL point DOC: Anthologie du</span></div> <div class="txt" style="position:absolute; left:333px; top:389px;"><span class="p9f7" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">développement au Sahel</span><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">. CD-ROM produced by</span></div> <div class="txt" style="position:absolute; left:333px; top:400px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">UNESCO, Paris, France.</span></div> <div class="txt" style="position:absolute; left:320px; top:414px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">10. UNU (1998) </span><span class="p9f7" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Collection on critical global issues. </span><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">CD-</span></div> <div class="txt" style="position:absolute; left:333px; top:425px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">ROM produced by the United Nations University</span></div> <div class="txt" style="position:absolute; left:333px; top:437px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Press, Tokyo, Japan.</span></div> <div class="txt" style="position:absolute; left:320px; top:451px;"><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">11. Witten, I.H., Moffat, A. and Bell, T. (1999) </span><span class="p9f7" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Managing</span></div> <div class="txt" style="position:absolute; left:333px; top:462px;"><span class="p9f7" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">Gigabytes: compressing and indexing documents and</span></div> <div class="txt" style="position:absolute; left:333px; top:473px;"><span class="p9f7" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">images</span><span class="p9f1" style="font-size:9px;vertical-align:baseline;color:rgba(0,0,0,1);">, Morgan Kaufmann, second edition.</span></div> </div> </div>