1 <chapter id="introduction">
2 <title>Introduction</title>
4 <section id="overview">
5 <title>Overview</title>
7 &zebra; is a free, fast, friendly information management system. It can
8 index records in &acro.xml;/&acro.sgml;, &acro.marc;, e-mail archives and many other
9 formats, and quickly find them using a combination of boolean
10 searching and relevance ranking. Search-and-retrieve applications can
11 be written using &acro.api;s in a wide variety of languages, communicating
12 with the &zebra; server using industry-standard information-retrieval
13 protocols or web services.
16 &zebra; is licensed Open Source, and can be
17 deployed by anyone for any purpose without license fees. The C source
18 code is open to anybody to read and change under the GPL license.
21 &zebra; is a networked component which acts as a
22 reliable &acro.z3950; server
23 for both record/document search, presentation, insert, update and
24 delete operations. In addition, it understands the &acro.sru; family of
25 webservices, which exist in &acro.rest; &acro.get;/&acro.post; and truly
29 &zebra; is available as MS Windows 2003 Server (32 bit) self-extracting
30 package as well as GNU/Debian Linux (32 bit and 64 bit) precompiled
31 packages. It has been deployed successfully on other Unix systems,
32 including Sun Sparc, HP Unix, and many variants of Linux and BSD
36 <ulink url="http://www.indexdata.com/zebra/">http://www.indexdata.com/zebra/</ulink>
37 <ulink url="http://ftp.indexdata.dk/pub/zebra/win32/">http://ftp.indexdata.dk/pub/zebra/win32/</ulink>
38 <ulink url="http://ftp.indexdata.dk/pub/zebra/debian/">http://ftp.indexdata.dk/pub/zebra/debian/</ulink>
42 <ulink url="http://indexdata.dk/zebra/">&zebra;</ulink>
43 is a high-performance, general-purpose structured text
44 indexing and retrieval engine. It reads records in a
45 variety of input formats (e.g. email, &acro.xml;, &acro.marc;) and provides access
46 to them through a powerful combination of boolean search
47 expressions and relevance-ranked free-text queries.
51 &zebra; supports large databases (tens of millions of records,
52 tens of gigabytes of data). It allows safe, incremental
53 database updates on live systems. Because &zebra; supports
54 the industry-standard information retrieval protocol, &acro.z3950;,
55 you can search &zebra; databases using an enormous variety of
56 programs and toolkits, both commercial and free, which understand
57 this protocol. Application libraries are available to allow
58 bespoke clients to be written in Perl, C, C++, Java, Tcl, Visual
59 Basic, Python, &acro.php; and more - see the
60 <ulink url="&url.zoom;">&acro.zoom; web site</ulink>
61 for more information on some of these client toolkits.
65 This document is an introduction to the &zebra; system. It explains
66 how to compile the software, how to prepare your first database,
67 and how to configure the server to give you the
68 functionality that you need.
72 <section id="features">
73 <title>&zebra; Features Overview</title>
75 <section id="features-document">
76 <title>&zebra; Document Model</title>
78 <table id="table-features-document" frame="top">
79 <title>&zebra; document model</title>
81 <colspec colwidth="1*" colname="feature"/>
82 <colspec colwidth="1*" colname="availability"/>
83 <colspec colwidth="3*" colname="notes"/>
84 <colspec colwidth="2*" colname="references"/>
87 <entry>Feature</entry>
88 <entry>Availability</entry>
90 <entry>Reference</entry>
95 <entry>Complex semi-structured Documents</entry>
96 <entry>&acro.xml; and &acro.grs1; Documents</entry>
97 <entry>Both &acro.xml; and &acro.grs1; documents exhibit a &acro.dom; like internal
98 representation allowing for complex indexing and display rules</entry>
99 <entry><xref linkend="record-model-alvisxslt"/> and
100 <xref linkend="grs"/></entry>
103 <entry>Input document formats</entry>
104 <entry>&acro.xml;, &acro.sgml;, Text, ISO2709 (&acro.marc;)</entry>
106 A system of input filters driven by
107 regular expressions allows most ASCII-based
108 data formats to be easily processed.
109 &acro.sgml;, &acro.xml;, ISO2709 (&acro.marc;), and raw text are also
111 <entry><xref linkend="componentmodules"/></entry>
114 <entry>Document storage</entry>
115 <entry>Index-only, Key storage, Document storage</entry>
116 <entry>Data can be, and usually is, imported
117 into &zebra;'s own storage, but &zebra; can also refer to
118 external files, building and maintaining indexes of "live"
128 <section id="features-search">
129 <title>&zebra; Search Features</title>
131 <table id="table-features-search" frame="top">
132 <title>&zebra; search functionality</title>
134 <colspec colwidth="1*" colname="feature"/>
135 <colspec colwidth="1*" colname="availability"/>
136 <colspec colwidth="3*" colname="notes"/>
137 <colspec colwidth="2*" colname="references"/>
140 <entry>Feature</entry>
141 <entry>Availability</entry>
143 <entry>Reference</entry>
148 <entry>Query languages</entry>
149 <entry>&acro.cql; and &acro.rpn;/&acro.pqf;</entry>
150 <entry>The type-1 Reverse Polish Notation (&acro.rpn;)
151 and its textual representation Prefix Query Format (&acro.pqf;) are
152 supported. The Common Query Language (&acro.cql;) can be configured as
153 a mapping from &acro.cql; to &acro.rpn;/&acro.pqf;</entry>
154 <entry><xref linkend="querymodel-query-languages-pqf"/> and
155 <xref linkend="querymodel-cql-to-pqf"/></entry>
158 <entry>Complex boolean query tree</entry>
159 <entry>&acro.cql; and &acro.rpn;/&acro.pqf;</entry>
160 <entry>Both &acro.cql; and &acro.rpn;/&acro.pqf; allow atomic query parts (&acro.apt;) to
161 be combined into complex boolean query trees</entry>
162 <entry><xref linkend="querymodel-rpn-tree"/></entry>
165 <entry>Field search</entry>
166 <entry>user defined</entry>
167 <entry>Atomic query parts (&acro.apt;) are either general, or
168 directed at user-specified document fields
170 <entry><xref linkend="querymodel-atomic-queries"/>,
171 <xref linkend="querymodel-use-string"/>,
172 <xref linkend="querymodel-bib1-use"/>, and
173 <xref linkend="querymodel-idxpath-use"/></entry>
176 <entry>Data normalization</entry>
177 <entry>user defined</entry>
178 <entry>Data normalization, text tokenization and character
179 mappings can be applied during indexing and searching</entry>
180 <entry><xref linkend="fields-and-charsets"/></entry>
183 <entry>Predefined field types</entry>
184 <entry>user defined</entry>
185 <entry>Data fields can be indexed as phrase, as into word
186 tokenized text, as numeric values, URLs, dates, and raw binary
188 <entry><xref linkend="character-map-files"/> and
189 <xref linkend="querymodel-pqf-apt-mapping-structuretype"/>
193 <entry>Regular expression matching</entry>
194 <entry>available</entry>
195 <entry>Full regular expression matching and "approximate
196 matching" (e.g. spelling mistake corrections) are handled.</entry>
197 <entry><xref linkend="querymodel-regular"/></entry>
200 <entry>Term truncation</entry>
201 <entry>left, right, left-and-right</entry>
202 <entry>The truncation attribute specifies whether variations of
203 one or more characters are allowed between search term and hit
204 terms, or not. Using non-default truncation attributes will
205 broaden the document hit set of a search query.</entry>
206 <entry><xref linkend="querymodel-bib1-truncation"/></entry>
209 <entry>Fuzzy searches</entry>
210 <entry>Spelling correction</entry>
211 <entry>In addition, fuzzy searches are implemented, where one
212 spelling mistake in search terms is matched</entry>
213 <entry><xref linkend="querymodel-bib1-truncation"/></entry>
220 <section id="features-scan">
221 <title>&zebra; Index Scanning</title>
223 <table id="table-features-scan" frame="top">
224 <title>&zebra; index scanning</title>
226 <colspec colwidth="1*" colname="feature"/>
227 <colspec colwidth="1*" colname="availability"/>
228 <colspec colwidth="3*" colname="notes"/>
229 <colspec colwidth="2*" colname="references"/>
232 <entry>Feature</entry>
233 <entry>Availability</entry>
235 <entry>Reference</entry>
241 <entry>term suggestions</entry>
242 <entry><literal>Scan</literal> on a given named index returns all the
243 indexed terms in lexicographical order near the given start
244 term. This can be used to create drop-down menus and search
246 <entry><xref linkend="querymodel-operation-type-scan"/> and
247 <xref linkend="querymodel-atomic-queries"/>
251 <entry>Facetted browsing</entry>
252 <entry>available</entry>
253 <entry>Zebra 2.1 and allows retrieval of facets for
256 <entry><xref linkend="querymodel-zebra-attr-scan"/></entry>
259 <entry>Drill-down or refine-search</entry>
260 <entry>partially</entry>
261 <entry>scanning in result sets can be used to implement
262 drill-down in search clients</entry>
263 <entry><xref linkend="querymodel-zebra-attr-scan"/></entry>
270 <section id="features-presentation">
271 <title>&zebra; Document Presentation</title>
273 <table id="table-features-presentation" frame="top">
274 <title>&zebra; document presentation</title>
276 <colspec colwidth="1*" colname="feature"/>
277 <colspec colwidth="1*" colname="availability"/>
278 <colspec colwidth="3*" colname="notes"/>
279 <colspec colwidth="2*" colname="references"/>
282 <entry>Feature</entry>
283 <entry>Availability</entry>
285 <entry>Reference</entry>
290 <entry>Hit count</entry>
292 <entry>Search results include at any time the total hit count of a given
293 query, either exact computed, or approximative, in case that the
294 hit count exceeds a possible pre-defined hit set truncation
297 <xref linkend="querymodel-zebra-local-attr-limit"/> and
298 <xref linkend="zebra-cfg"/>
302 <entry>Paged result sets</entry>
304 <entry>Paging of search requests and present/display request
305 can return any successive number of records from any start
306 position in the hit set, i.e. it is trivial to provide search
307 results in successive pages of any size.</entry>
311 <entry>&acro.xml; document transformations</entry>
312 <entry>&acro.xslt; based</entry>
313 <entry> Record presentation can be performed in many
314 pre-defined &acro.xml; data
315 formats, where the original &acro.xml; records are on-the-fly transformed
316 through any preconfigured &acro.xslt; transformation. It is therefore
317 trivial to present records in short/full &acro.xml; views, transforming to
318 RSS, Dublin Core, or other &acro.xml; based data formats, or transform
319 records to XHTML snippets ready for inserting in XHTML pages.</entry>
321 <xref linkend="record-model-alvisxslt-elementset"/></entry>
324 <entry>Binary record transformations</entry>
325 <entry>&acro.marc;, &acro.usmarc;, &acro.marc21; and &acro.marcxml;</entry>
326 <entry>post-filter record transformations</entry>
330 <entry>Record Syntaxes</entry>
332 <entry> Multiple record syntaxes
333 for data retrieval: &acro.grs1;, &acro.sutrs;,
334 &acro.xml;, ISO2709 (&acro.marc;), etc. Records can be mapped between
335 record syntaxes and schemas on the fly.</entry>
339 <entry>&zebra; internal metadata</entry>
341 <entry> &zebra; internal document metadata can be fetched in
342 &acro.sutrs; and &acro.xml; record syntaxes. Those are useful in client
343 applications.</entry>
344 <entry><xref linkend="special-retrieval"/></entry>
347 <entry>&zebra; internal raw record data</entry>
349 <entry> &zebra; internal raw, binary record data can be fetched in
350 &acro.sutrs; and &acro.xml; record syntaxes, leveraging %zebra; to a
351 binary storage system</entry>
352 <entry><xref linkend="special-retrieval"/></entry>
355 <entry>&zebra; internal record field data</entry>
357 <entry> &zebra; internal record field data can be fetched in
358 &acro.sutrs; and &acro.xml; record syntaxes. This makes very fast minimal
359 record data displays possible.</entry>
360 <entry><xref linkend="special-retrieval"/></entry>
367 <section id="features-sort-rank">
368 <title>&zebra; Sorting and Ranking</title>
370 <table id="table-features-sort-rank" frame="top">
371 <title>&zebra; sorting and ranking</title>
373 <colspec colwidth="1*" colname="feature"/>
374 <colspec colwidth="1*" colname="availability"/>
375 <colspec colwidth="3*" colname="notes"/>
376 <colspec colwidth="2*" colname="references"/>
379 <entry>Feature</entry>
380 <entry>Availability</entry>
382 <entry>Reference</entry>
388 <entry>numeric, lexicographic</entry>
389 <entry>Sorting on the basis of alpha-numeric and numeric data
390 is supported. Alphanumeric sorts can be configured for
391 different data encodings and locales for European languages.</entry>
392 <entry><xref linkend="administration-ranking-sorting"/> and
393 <xref linkend="querymodel-zebra-attr-sorting"/></entry>
396 <entry>Combined sorting</entry>
398 <entry>Sorting on the basis of combined sorts  e.g. combinations of
399 ascending/descending sorts of lexicographical/numeric/date field data
401 <entry><xref linkend="administration-ranking-sorting"/></entry>
404 <entry>Relevance ranking</entry>
405 <entry>TF-IDF like</entry>
406 <entry>Relevance-ranking of free-text queries is supported
407 using a TF-IDF like algorithm.</entry>
408 <entry><xref linkend="administration-ranking-dynamic"/></entry>
411 <entry>Static pre-ranking</entry>
413 <entry>Enables pre-index time ranking of documents where hit
414 lists are ordered first by ascending static rank, then by
415 ascending document ID.</entry>
416 <entry><xref linkend="administration-ranking-static"/></entry>
424 <section id="features-updates">
425 <title>&zebra; Live Updates</title>
428 <table id="table-features-updates" frame="top">
429 <title>&zebra; live updates</title>
431 <colspec colwidth="1*" colname="feature"/>
432 <colspec colwidth="1*" colname="availability"/>
433 <colspec colwidth="3*" colname="notes"/>
434 <colspec colwidth="2*" colname="references"/>
437 <entry>Feature</entry>
438 <entry>Availability</entry>
440 <entry>Reference</entry>
445 <entry>Incremental and batch updates</entry>
447 <entry>It is possible to schedule record inserts/updates/deletes in any
448 quantity, from single individual handled records to batch updates
449 in strikes of any size, as well as total re-indexing of all records
450 from file system. </entry>
451 <entry><xref linkend="zebraidx"/></entry>
454 <entry>Remote updates</entry>
455 <entry>&acro.z3950; extended services</entry>
456 <entry>Updates can be performed from remote locations using the
457 &acro.z3950; extended services. Access to extended services can be
458 login-password protected.</entry>
459 <entry><xref linkend="administration-extended-services"/> and
460 <xref linkend="zebra-cfg"/></entry>
463 <entry>Live updates</entry>
464 <entry>transaction based</entry>
465 <entry> Data updates are transaction based and can be performed
466 on running &zebra; systems. Full searchability is preserved
467 during life data update due to use of shadow disk areas for
468 update operations. Multiple update transactions at the same
469 time are lined up, to be performed one after each other. Data
470 integrity is preserved.</entry>
471 <entry><xref linkend="shadow-registers"/></entry>
478 <section id="features-protocol">
479 <title>&zebra; Networked Protocols</title>
481 <table id="table-features-protocol" frame="top">
482 <title>&zebra; networked protocols</title>
484 <colspec colwidth="1*" colname="feature"/>
485 <colspec colwidth="1*" colname="availability"/>
486 <colspec colwidth="3*" colname="notes"/>
487 <colspec colwidth="2*" colname="references"/>
490 <entry>Feature</entry>
491 <entry>Availability</entry>
493 <entry>Reference</entry>
498 <entry>Fundamental operations</entry>
499 <entry>&acro.z3950;/&acro.sru; <literal>explain</literal>,
500 <literal>search</literal>, <literal>scan</literal>, and
501 <literal>update</literal></entry>
503 <entry><xref linkend="querymodel-operation-types"/></entry>
506 <entry>&acro.z3950; protocol support</entry>
508 <entry> Protocol facilities supported are:
509 <literal>init</literal>, <literal>search</literal>,
510 <literal>present</literal> (retrieval),
511 Segmentation (support for very large records),
512 <literal>delete</literal>, <literal>scan</literal>
513 (index browsing), <literal>sort</literal>,
514 <literal>close</literal> and support for the <literal>update</literal>
515 Extended Service to add or replace an existing &acro.xml;
516 record. Piggy-backed presents are honored in the search
517 request. Named result sets are supported.</entry>
518 <entry><xref linkend="protocol-support"/></entry>
521 <entry>Web Service support</entry>
522 <entry>&acro.sru;</entry>
523 <entry> The protocol operations <literal>explain</literal>,
524 <literal>searchRetrieve</literal> and <literal>scan</literal>
525 are supported. <ulink url="&url.cql;">&acro.cql;</ulink> to internal
526 query model &acro.rpn;
527 conversion is supported. Extended RPN queries
528 for search/retrieve and scan are supported.</entry>
529 <entry><xref linkend="zebrasrv-sru-support"/></entry>
536 <section id="features-scalability">
537 <title>&zebra; Data Size and Scalability</title>
539 <table id="table-features-scalability" frame="top">
540 <title>&zebra; data size and scalability</title>
542 <colspec colwidth="1*" colname="feature"/>
543 <colspec colwidth="1*" colname="availability"/>
544 <colspec colwidth="3*" colname="notes"/>
545 <colspec colwidth="2*" colname="references"/>
548 <entry>Feature</entry>
549 <entry>Availability</entry>
551 <entry>Reference</entry>
556 <entry>No of records</entry>
557 <entry>40-60 million</entry>
562 <entry>Data size</entry>
563 <entry>100 GB of record data</entry>
564 <entry>&zebra; based applications have successfully indexed up
565 to 100 GB of record data</entry>
569 <entry>Scale out</entry>
570 <entry>multiple discs</entry>
575 <entry>Performance</entry>
576 <entry><literal>O(n * log N)</literal></entry>
577 <entry> &zebra; query speed and performance is affected roughly by
578 <literal>O(log N)</literal>,
579 where <literal>N</literal> is the total database size, and by
580 <literal>O(n)</literal>, where <literal>n</literal> is the
581 specific query hit set size.</entry>
585 <entry>Average search times</entry>
587 <entry> Even on very large size databases hit rates of 20 queries per
588 seconds with average query answering time of 1 second are possible,
589 provided that the boolean queries are constructed sufficiently
590 precise to result in hit sets of the order of 1000 to 5.000
595 <entry>Large databases</entry>
596 <entry>64 bit file pointers</entry>
597 <entry>64 file pointers assure that register files can extend
598 the 2 GB limit. Logical files can be
599 automatically partitioned over multiple disks, thus allowing for
600 large databases.</entry>
608 <section id="features-platforms">
609 <title>&zebra; Supported Platforms</title>
611 <table id="table-features-platforms" frame="top">
612 <title>&zebra; supported platforms</title>
614 <colspec colwidth="1*" colname="feature"/>
615 <colspec colwidth="1*" colname="availability"/>
616 <colspec colwidth="3*" colname="notes"/>
617 <colspec colwidth="2*" colname="references"/>
620 <entry>Feature</entry>
621 <entry>Availability</entry>
623 <entry>Reference</entry>
630 <entry>GNU Linux (32 and 64bit), journaling Reiser or (better)
632 on disks. NFS file systems are not supported.
633 GNU/Debian Linux packages are available</entry>
634 <entry><xref linkend="installation-debian"/></entry>
638 <entry>tar-ball</entry>
639 <entry>&zebra; is written in portable C, so it runs on most
641 Usual tar-ball install possible on many major Unix systems</entry>
642 <entry><xref linkend="installation-unix"/></entry>
645 <entry>Windows</entry>
646 <entry>NT/2000/2003/XP</entry>
647 <entry>&zebra; runs as well on Windows (NT/2000/2003/XP).
648 Windows installer packages available</entry>
649 <entry><xref linkend="installation-win32"/></entry>
659 <section id="introduction-apps">
660 <title>References and &zebra; based Applications</title>
662 &zebra; has been deployed in numerous applications, in both the
663 academic and commercial worlds, in application domains as diverse
664 as bibliographic catalogues, Geo-spatial information, structured
665 vocabulary browsing, government information locators, civic
666 information systems, environmental observations, museum information
670 Notable applications include the following:
674 <section id="koha-ils">
675 <title>Koha free open-source ILS</title>
677 <ulink url="http://www.koha.org/">Koha</ulink> is a full-featured
678 open-source ILS, initially developed in
679 New Zealand by Katipo Communications Ltd, and first deployed in
680 January of 2000 for Horowhenua Library Trust. It is currently
681 maintained by a team of software providers and library technology
682 staff from around the globe.
685 <ulink url="http://liblime.com/">LibLime</ulink>,
686 a company that is marketing and supporting Koha, adds in
687 the new release of Koha 3.0 the &zebra;
688 database server to drive its bibliographic database.
691 In early 2005, the Koha project development team began looking at
692 ways to improve &acro.marc; support and overcome scalability limitations
693 in the Koha 2.x series. After extensive evaluations of the best
694 of the Open Source textual database engines - including MySQL
695 full-text searching, PostgreSQL, Lucene and Plucene - the team
699 "&zebra; completely eliminates scalability limitations, because it
700 can support tens of millions of records." explained Joshua
701 Ferraro, LibLime's Technology President and Koha's Project
702 Release Manager. "Our performance tests showed search results in
703 under a second for databases with over 5 million records on a
704 modest i386 900Mhz test server."
707 "&zebra; also includes support for true boolean search expressions
708 and relevance-ranked free-text queries, both of which the Koha
709 2.x series lack. &zebra; also supports incremental and safe
710 database updates, which allow on-the-fly record
711 management. Finally, since &zebra; has at its heart the &acro.z3950;
712 protocol, it greatly improves Koha's support for that critical
716 Although the bibliographic database will be moved to &zebra;, Koha
717 3.0 will continue to use a relational SQL-based database design
718 for the 'factual' database. "Relational database managers have
719 their strengths, in spite of their inability to handle large
720 numbers of bibliographic records efficiently," summed up Ferraro,
721 "We're taking the best from both worlds in our redesigned Koha
725 See also LibLime's newsletter article
726 <ulink url="http://www.liblime.com/newsletter/2006/01/features/koha-earns-its-stripes/">
727 Koha Earns its Stripes</ulink>.
732 <section id="kete-dom">
733 <title>Kete Open Source Digital Library and Archiving software</title>
735 <ulink url="http://kete.net.nz/">Kete</ulink> is a digital object
736 management repository, initially developed in
737 New Zealand. Initial development has
738 been a partnership between the Horowhenua Library Trust and
739 Katipo Communications Ltd. funded as part of the Community
740 Partnership Fund in 2006.
741 Kete is purpose built
742 software to enable communities to build their own digital
743 libraries, archives and repositories.
746 It is based on Ruby-on-Rails and MySQL, and integrates the &zebra; server
747 and the &yaz; toolkit for indexing and retrieval of it's content.
748 Zebra is run as separate computer process from the Kete
752 url="http://kete.net.nz/documentation/topics/show/139-managing-zebra">manages
756 Why does Kete wants to use Zebra?? Speed, Scalability and easy
757 integration with Koha. Read their
759 url="http://kete.net.nz/blog/topics/show/44-who-what-why-when-answering-some-of-the-niggly-development-questions">detailed
760 reasoning here.</ulink>
764 <section id="reindex-ils">
765 <title>ReIndex.Net web based ILS</title>
767 <ulink url="http://www.reindex.net/index.php?lang=en">Reindex.net</ulink>
768 is a netbased library service offering all
769 traditional functions on a very high level plus many new
770 services. Reindex.net is a comprehensive and powerful WEB system
771 based on standards such as &acro.xml; and &acro.z3950;.
772 updates. Reindex supports &acro.marc21;, dan&acro.marc; eller Dublin Core with
776 Reindex.net runs on GNU/Debian Linux with &zebra; and Simpleserver
778 Data for bibliographic data. The relational database system
779 Sybase 9 &acro.xml; is used for
781 Internally &acro.marcxml; is used for bibliographical records. Update
782 utilizes &acro.z3950; extended services.
786 <section id="dads-article-database">
787 <title>DADS - the DTV Article Database
790 DADS is a huge database of more than ten million records, totalling
791 over ten gigabytes of data. The records are metadata about academic
792 journal articles, primarily scientific; about 10% of these
793 metadata records link to the full text of the articles they
794 describe, a body of about a terabyte of information (although the
795 full text is not indexed.)
798 It allows students and researchers at DTU (Danmarks Tekniske
799 Universitet, the Technical College of Denmark) to find and order
800 articles from multiple databases in a single query. The database
801 contains literature on all engineering subjects. It's available
802 on-line through a web gateway, though currently only to registered
806 More information can be found at
807 <ulink url="http://www.dtic.dtu.dk/"/> and
808 <ulink url="http://dads.dtv.dk"/>
813 <title>ULS (Union List of Serials)</title>
816 has created a union catalogue for the periodicals of the
817 twenty-one constituent libraries of the University of London and
818 the University of Westminster
819 (<ulink url="http://www.m25lib.ac.uk/ULS/"/>).
820 They have achieved this using an
821 unusual architecture, which they describe as a
822 ``non-distributed virtual union catalogue''.
825 The member libraries send in data files representing their
826 periodicals, including both brief bibliographic data and summary
827 holdings. Then 21 individual &acro.z3950; targets are created, each
828 using &zebra;, and all mounted on the single hardware server.
829 The live service provides a web gateway allowing &acro.z3950; searching
830 of all of the targets or a selection of them. &zebra;'s small
831 footprint allows a relatively modest system to comfortably host
835 More information can be found at
836 <ulink url="http://www.m25lib.ac.uk/ULS/"/>
840 <section id="various-web-indexes">
841 <title>Various web indexes</title>
843 &zebra; has been used by a variety of institutions to construct
844 indexes of large web sites, typically in the region of tens of
845 millions of pages. In this role, it functions somewhat similarly
846 to the engine of Google or AltaVista, but for a selected intranet
847 or a subset of the whole Web.
850 For example, Liverpool University's web-search facility (see on
852 <ulink url="http://www.liv.ac.uk/"/>
853 and many sub-pages) works by relevance-searching a &zebra; database
854 which is populated by the Harvest-NG web-crawling software.
857 For more information on Liverpool university's intranet search
858 architecture, contact John Gilbertson
859 <email>jgilbert@liverpool.ac.uk</email>
863 has recently modified the Harvest web indexer to use &zebra; as
864 its native repository engine. His comments on the switch over
865 from the old engine are revealing:
868 The first results after some testing with &zebra; are very
869 promising. The tests were done with around 220,000 SOIF files,
870 which occupies 1.6GB of disk space.
873 Building the index from scratch takes around one hour with &zebra;
874 where [old-engine] needs around five hours. While [old-engine]
875 blocks search requests when updating its index, &zebra; can still
876 answer search requests.
878 &zebra; supports incremental indexing which will speed up indexing
882 While the search time of [old-engine] varies from some seconds
883 to some minutes depending how expensive the query is, &zebra;
884 usually takes around one to three seconds, even for expensive
887 &zebra; can search more than 100 times faster than [old-engine]
888 and can process multiple search requests simultaneously
891 I am very happy to see such nice software available under GPL.
898 <section id="introduction-support">
899 <title>Support</title>
901 You can get support for &zebra; from at least three sources.
904 First, there's the &zebra; web site at
905 <ulink url="&url.idzebra;"/>,
906 which always has the most recent version available for download.
907 If you have a problem with &zebra;, the first thing to do is see
908 whether it's fixed in the current release.
911 Second, there's the &zebra; mailing list. Its home page at
912 <ulink url="&url.idzebra.mailinglist;"/>
913 includes a complete archive of all messages that have ever been
914 posted on the list. The &zebra; mailing list is used both for
915 announcements from the authors (new
916 releases, bug fixes, etc.) and general discussion. You are welcome
917 to seek support there. Join by filling the form on the list home page.
921 <!-- Keep this comment at the end of the file
926 sgml-minimize-attributes:nil
927 sgml-always-quote-attributes:t
930 sgml-parent-document: "idzebra.xml"
931 sgml-local-catalogs: nil
932 sgml-namecase-general:t