Introduction to XML: A Librarians Perspective

1
Introduction to XMLA Librarians Perspective

• Delphine KhannaRutgers UniversityPalinet,
  July/August 1999

2
Overview of the workshop

• What is XML? How does it work?
• Why XML? What is it going to change?
• Overview of some XML-related standards.
• XML in libraries (standards projects).
• Practical skills
• Creating an XML document,
• Creating an XSL style sheet,
• Work with MS Internet Explorer 5.0.

3
Workshop Web site

• http//scc01.rutgers.edu/ceth/intromat/xml/
• Contents
• This slide presentation,
• XML Samples used in this workshop,
• List of useful Web links and other resources.

4
A First Look at XML
5
Basics

• Simplified definition XML is a kind of
  super-HTML where you can define your own tags.

6
Term Clarification

• XML can be called a
• encoding format,
• language,
• standard.
• We will prefer standard.

7
The XML Family A whole family of standards

• XML,
• XSL,
• XLINK, XPOINTER,
• Namespaces,
• RDF,
• XML Schemas,
• DOM,
• and more

8
XML Who? When?

• XML family developed by W3C.
• Very recent
• XML 1.0 February 1998.
• Namespaces January 1999.
• RDF February 1999.
• XLINK, XPOINTER, XSL, XML Schemas still working
  drafts.

9
XML To develop the next generation of Web
applications

• People want to do more sophisticated things with
  the Web.
• HTML is too limited for that.
• Need for a more powerful language XML.

10
XML Hype 2 myths

• XML will replace everything.
• (HTML, back-end relational databases, etc.)
• XML is completely different from Web technologies
  we had before.

11
Why is XML better?
12
Lets look at a typical HTML document

• ltBODYgt
• ltH1gtLines Written in Early Springlt/H1gt
• ltH2gtWilliam Wordsworthlt/H2gt
• ltPgtI heard a thousand blended notes,ltBRgt
• While in grove I sate reclined,ltBRgt
• In that sweet mood when pleasant
  thoughtsltBRgt
• Bring sad thoughts to the mind.lt/Pgt
• ltPgtTo her fair works did nature linkltBRgt
• The human soul that through me ranltBRgt
• And much it griev'd me my heart to
  thinkltBRgt
• What man has made of man.lt/Pgt
• lt/BODYgt

13
What is the problem?

• To do more fancy things with documents
• need to make their logical structure explicit.
• Otherwise software applications
• do not know what is what,
• do not have any handle over documents.

14
Why XML is better Overall

• HTML
• Encoding too vague and messy.
• Logical structure is not clearly encoded.
• XML
• Allows us to create clean structured documents,
  where logical structure of document is totally
  explicit.

15
The same document in XML

• lt?xml version"1.0"?gt
• ltPOEMgt
• ltTITLEgtLines Written in Early Springlt/TITLEgt
• ltAUTHORgtltFIRSTNAMEgtWilliamlt/FIRSTNAMEgt
  ltLASTNAMEgtWordsworthlt/LASTNAMEgtlt/AUTHOR
  gt
• ltSTANZAgt
• ltLINE N1gtI heard a thousand blended
  notes,lt/LINEgt
• ltLINE N2gtWhile in grove I sate
  reclined,lt/LINEgt
• ltLINE N3gtIn that sweet mood when
  pleasant thoughtslt/LINEgt
• ltLINE N4gtBring sad thoughts to the
  mind.lt/LINEgt
• lt/STANZAgt
• ltSTANZAgt
• ltLINE N5gtTo her fair works did nature
  linklt/LINEgt
• ltLINE N6gtThe human soul that through
  me ranlt/LINEgt
• ltLINE N7gtAnd much it griev'd me my
  heart to thinklt/LINEgt
• ltLINE N8gtWhat man has made of
  man.lt/LINEgt
• lt/STANZAgt

16
Why XML is better Reason 1

• HTML One single fixed tag set
• ltH1gt, ltH2gt, ltPgt, ltIMGgt, etc.
• XML You can define your own tag set
• ltPOEMgt, ltSTANZAgt, ltLINE N18gt.
• ltPgt, ltABSTRACTgt, ltFOOTNOTEgt, ltBIBL_ENTRYgt.
• gt Possible to describe the logical structure
  exactly.

17
Why XML is better Reason 2

• HTML Lack of syntax controlltAgtltPgtHellolt/Agtlt/Pgtlt
  Pgtis considered OK.
• XML Documents have to be at least
  well-formedltPgtltAgtHellolt/Agtlt/Pgtis the only
  form acceptable.
• gt Code much cleaner.

18
Why XML is better Reason 3

• HTML Logical structure and display are mixed up
• ltPgt, ltH1gt, ltH2gt.
• This text is ltFONT COLORbluegtimportantlt/FONTgt.
  
• XML Clear distinction between logical structure
  and display
• This text is ltEMPHgtimportantlt/EMPHgt.
• ltEMPHgt ltFONT COLORbluegt
• gt Code much cleaner.

19
By the way, HTML is not that bad

• HTML
• Really simple Attractive to basic users.
• Works fine for basic Web pages.
• XML
• Clearly more complexWill scare off basic users.
• Probably an overkill for basic Web pages.

20
What will XML change?Or, why do we need to
make the logical structure explicit?
21
Different displays for different output devices

• Regular computer screens,
• Pocket computers, Palm Pilot,
• WebTV,
• Audio (visually-impaired, cars),
• Braille,
• Print.

22
Term Clarification

• The Web based on a client/server architecture.

23
Server-side Databases should speak to each other

• A very successful model
• relational databases on the server side.
• Next step data integration.
• Example 1 An online bookstore,
• Example 2 Medical records,
• Example 3 An index that knows which journals
  are available in the library.

24
XML representing structured data

• If XML can represent structured text,it can also
  represent structured data.
• XML is also very good at representing mixed data
  seamlessly.

25
XML for Interchange Example of a converted R-DB
record

• lt?xml version"1.0?gt
• ltPACKAGEgt
• ltIDgt33456lt/IDgt
• ltCATEGORYgtNext day deliverylt/CATEGORYgt
• ltSHIPPING_COSTgt15lt/ SHIPPING_ COSTgt
• ltLOC_DEPARTUREgtNew York Citylt/LOC_DEPARTUREgt
• ltLOC_ARRIVALgtPittsburghlt/LOC_ARRIVALgt
• ltDATE_DEPARTUREgt07/30/1999lt/ DATE_DEPARTUREgt
• ltDATE_ARRIVALgt07/31/1999lt/ DATE_ARRIVALgt
• ltPACKAGEgt

26
Client-side The Web more than an online
fax-machine

• Web-browsers thin clients
• They just display documents.
• Clients can do more
• Client workstation has a lot of unused power,
• Less strain on the network and on the server,
• Example Viewing and sorting of a medical record.

27
Client-side The Web more than an online
fax-machine (2)

• Clients can do more
• Personalized and sophisticated processing
  possible.
• Processing possibly provided by 3rd-party client
  applications.
• Example Bibliography manager.

28
XML The nitty-gritty details
29
Term Clarification

• Element,
• Tag (opening tag / closing tag / delimiter),
• Element content,
• Attribute (name / value).
• Example
• ltAUTHOR TYPEnovelistgtJohn Smithlt/AUTHORgt

30
Differences in Syntax between XML and HTML

• XML Declarationlt?xml version"1.0?gt
• Every opening tag must have a closing tag.
• Empty tags have a different syntax ltBR/gt
• Tags are case sensitive ltSTANZAgt different from
  ltstanzagt

31
2-Level Syntax Control

• XML documents can be
• Well-formed,
• Valid.

32
Syntax Control Well-formed documents

• All XML documents must be well-formed.
• XML parsers check the well-formedness.
• Criteria of well-formedness
• Every opening tag must have a closing tag.
  Illegal ltPgtHello
• No overlapping elements Illegal
  ltDIVgtltPgtHellolt/DIVgtlt/Pgt
• One unique root element

33
Tree Representation

• POEM
• TITLE AUTHOR STANZA STANZA
• FIRSTNAME LASTNAME LINE LINE
  LINE LINE LINE LINE

34
Create your own XML document

• The cooking recipe document
• 1. Brainstorming on the structure of the
  document,
• 2. Creation of the document with a template.

35
Editing XML Documents

• Textpad Internet Explorer 5 as a parser.
• Caution IE5 comes with limitations and
  proprietary features.
• Alternative
• XML editor (e.g. Softquads XMetal).

36
To get started

• Create file in Textpad and load it in IE5.
• File extension xml.
• Save regularly and reload in IE5.
• Begin with
• lt?xml version"1.0"?gt
• ltRECIPEgt
• lt/ RECIPE gt

37
Document Type Definitions (DTD)
38
Document Type Definitions(DTD)

• Formal way of defining the tags used in a series
  of documents.
• A DTD
• specifies a list of tags,
• defines the relationships between these tags.
• Allows us to create consistency across a
  collection of documents (e.g., 5000 poems).

39
How does a DTD look like?

• lt!ELEMENT POEM (TITLE, AUTHOR, STANZA)gt
• lt!ELEMENT TITLE (PCDATA)gt
• lt!ELEMENT AUTHOR (FIRSTNAME, LASTNAME)gt
• lt!ELEMENT FIRSTNAME (PCDATA)gt
• lt!ELEMENT LASTNAME (PCDATA)gt
• lt!ELEMENT STANZA (LINE)gt
• lt!ELEMENT LINE (PCDATA)gt
• lt!ATTLIST LINE N CDATA REQUIREDgt

40
Creating a DTD

• Non-trivial task.
• Higher level of expertise needed than for using a
  DTD.
• In-depth knowledge of XML,
• In-depth knowledge of the type of documents being
  described.
• Preliminary Document Analysis.
• A DTD can be dozens of pages long.

41
Syntax ControlValid documents

• Higher level of control than well-formed
  documents.
• An XML document is valid if it conforms to its
  DTD.
• To validate an XML Document, it is necessary to
  declare the name and location of its DTD.

42
XML DTD declaration

• The DTD should be declared at the top of the XML
  document.
• Local file
• lt?xml version1.0 standaloneno?gt
• lt!DOCTYPE recipe SYSTEM poem.dtdgt
• URL
• lt?xml version1.0 standaloneno?gt
• lt!DOCTYPE recipe SYSTEM http//scc01.rutgers.
  edu/ceth/intromat/xml/samples/poem/poem.dtdgt

43
Validation with IE5

• When loading a documentThe IE5 parser does not
  validate it.
• Possible to validate a document through a script.
• Possible also to use a separate validating
  parser.
• For instance, the Scholarly Technology Groups
  XML parser at Brown U.
• (http//www.stg.brown.edu/service/xmlvalid/).
• Validating vs. non-validating parsers.

44
Validation Strategy

• For now, best model
• When creating documents use a validating parser.
• (like the Scholarly Technology Group's XML
  Parser)
• When users download them parser only checks if
  well formed.

45
Namespaces

• Need to use elements from several DTDs in the
  same document.
• Scheme to identify the source of each element.
• Special case Same element name used by 2 DTDs.

46
Namespace Example

• ltbook xmlnsbooks'urnloc.govbooks'
• xmlnsisbn'urlISBNhttp//www.isbn.o
  rg/isbndtdgt
• ltbookstitlegtCheaper by the
  Dozenlt/bookstitlegt
• ltisbnnumbergt1568491379lt/isbnnumbergt
• ltbooksnotesgtThis is funny
  booklt/booksnotesgt
• lt/bookgt
• Note Adapted from example in the Namespaces
  recommendation.

47
Namespace Example (2)Default Namespace

• ltbook xmlns'urnloc.govbooks'
• xmlnsisbn'urlISBNhttp//www.isbn.org
  /isbndtdgt
• lttitlegtCheaper by the Dozenlt/titlegt
• ltisbnnumbergt1568491379lt/isbnnumbergt
• ltnotesgtThis is funny booklt/notesgt
• lt/bookgt
• Note Adapted from example in the Namespaces
  recommendation.

48
More good things about XML
49
Positive side-effects of XML (1)

• XML fosters the development of community-based
  standards.
• Concept of 2-level standard very powerful
• XML universal,
• DTDs community-specific.
• Now developing a new standard amounts to writing
  a DTD.
• Much easier than starting from scratch.
• E.g., Xlit.

50
Positive side-effects of XML (2)

• Wide-spread standards are stronger than those
  used by a limited community(regardless of their
  intrinsic value).
• HL7 --gt XML.
• Easier to hire programmers.
• More documentation available.
• Actively maintained by very large base of people.

51
Positive side-effects of XML (3)

• A set of standards bundled together are stronger
  than an isolated one.
• Likely to appeal to more people (The Microsoft
  Office idea).
• The standards reinforce each others.

52
Stylesheet Languages for XML
53
Stylesheet Languages for XML

• Specify how to display logical elements.
• XML supports 2 stylesheet languages
• CSS
• Quite Limited,
• But eases transition HTML--gtXML.
• XSL
• Very powerful,
• Still a working draft.

54
Extensible Stylesheet Language (XSL)

• 2 Parts
• Transformations
• Transform the XML document (reorder, hide, add
  elements).
• Formatting Objects (FO)
• Attach formatting properties to XML elements.

55
XSL in IE 5.0

• Supports transformations but not the FO.
• Trick transform XML DTD-specific elements into
  HTML elements.
• Convenient because everybody knows HTML.

56
XSL-to-HTML Stylesheets Syntax

• Style Sheet Excerpt XML Document Excerpt
• ltxsltemplate matchBOOK"gt ltBOOKgt
• ltPgtltxslapply-templates/gtlt/Pgt
  ltAUTHORgtMary Brownlt/AUTHORgt
• lt/xsltemplategt ltTITLEgtEasy
  Cookinglt/TITLEgt
• lt/BOOKgt
• ltxsltemplate matchAUTHORgt ltBOOKgt
• ltBgt ltxslvalue-of/gtlt/Bgt
  ltAUTHORgtJohn Smithlt/AUTHORgt
• lt/xsltemplategt ltTITLEgt101
  Recipeslt/TITLEgt
• lt/BOOKgt
• ltxsltemplate matchTITLE"gt ltBOOKgt
• ltxslvalue-of/gt ltAUTHORgtSue
  Meyerlt/AUTHORgt
• lt/xsltemplategt ltTITLEgtItalian
  Cuisinelt/TITLEgt
• lt/BOOKgt
• HTML Output
• ltPgtltBgtMary Brownlt/Bgt Easy Cookinglt/Pgt
• ltPgtltBgtJohn Smithlt/Bgt 101 Recipeslt/Pgt
• ltPgtltBgtSue Meyerlt/Bgt Italian Cuisinelt/Pgt

57
Beginning of an XSL-to-HTML Stylesheet

• lt?xml version'1.0'?gt
• ltxslstylesheet xmlnsxsl"http//www.w3.org/TR/WD
  -xsl"gt
• ltxsltemplate match"/"gt
• ltxslapply-templates/gt
• lt/xsltemplategt
• ltxsltemplate match"POEM"gt
• ltHTMLgt
• ltBODYgt
• ltxslapply-templates/gt
• lt/BODYgt
• lt/HTMLgt
• lt/xsltemplategt
• lt/xslstylesheetgt

58
Example of XSL-to-HTML Stylesheet

• See poem.xsl at
• http//scc01.rutgers.edu/ceth/intromat/xml/samples
  /poem/poem.xsl

59
Declaring an XSL Stylesheet in an XML document

• Just after the XML declaration (and the DTD
  declaration if there is one).
• Local file
• lt?xml-stylesheet typetext/xsl
  hrefpoem.xsl?gt
• URL
• lt?xml-stylesheet typetext/xsl
  hrefhttp//scc01.rutg ers.edu/ceth/intromat/xml/
  samples/poem/poem.xsl ?gt

60
Creating your own Stylesheet

• The XSL-to-HTML recipe stylesheet
• XSL stylesheets can be tricky.
• Always use another stylesheet as a model.
• Name the file recipe.xsl.
• Make sure to declare it in the XML document.
• lt?xml-stylesheet type"text/xsl"
  hrefrecipe.xsl"?gt
• Always add one template at a time, and reload in
  IE5 to make sure it works.

61
Recipe Stylesheet Step 1

• lt?xml version"1.0"?gt
• ltxslstylesheet xmlnsxsl"http//www.w3.org/TR/WD
  -xsl"gt
• ltxsltemplate match"/"gt
• ltxslapply-templates/gt
• lt/xsltemplategt
• lt/xslstylesheetgt

62
Recipe Stylesheet Step 2

• ltxsltemplate match"RECIPE"gt
• ltHTMLgt
• ltBODY BGCOLOR"FFFFCC"gt
• ltxslapply-templates/gt
• lt/BODYgt
• lt/HTMLgt
• lt/xsltemplategt
• ltxsltemplate match"TITLE"gt
• ltH1gtltCENTERgtltFONT COLOR"red"gt
• ltxslvalue-of/gt
• lt/FONTgtlt/CENTERgtlt/H1gt
• lt/xsltemplategt

63
Recipe Stylesheet Step 3 and after

• For the rest of the stylesheet, see the sample
  recipe.xsl at
• http//scc01.rutgers.edu/ceth/intromat/xml/samples
  /recipe/recipe.xsl

64
Recipe Stylesheet Advanced

• Sorting the ingredients
• ltxsltemplate match"INGREDIENTLIST"gt
• ltHR/gt
• ltH2gtltFONT COLOR"red"gtIngredientslt/FONTgtlt/H2gt
  
• ltULgtltxslapply-templates select"INGREDIENT"
• order-by"PRODUCT/_at_AISLE
  PRODUCT"/gtlt/ULgt
• ltHR/gt
• lt/xsltemplategt

65
XML Formatting ObjectsExample

• ltxsltemplate matchtitlegt
• ltfoblock font-weightbold font-colorrgb(0,255
  ,255)
• font-size16ptgt
• ltxslapply-templates/gt
• lt/foblockgt
• lt/xsltemplategt
• Note Adapted from stylesheet created by Lynn
  Lobash.

66
Some Other XML-related Standards
67
Linking Standards

• HTML links
• Really primitive and limited.
• Linking standards for XML
• Much more powerful.
• 2 parts
• XLink (aka. XLL),
• XPointer (aka. XLP).
• Still working drafts.

68
XLink

• To define links to one or several documents.
• 2 types of links
• Simple,
• Extended.

69
XLink Simple link

• Example
• ltrelated_poem xmllinksimple inlinefalse
  hrefpoem1.xmlgtGo to related poemlt/related_poemgt
  
• Other attributes / Alternative values
• inline true, false (link to same document vs.
  outside).
• show replace, new, embed.
• actuate user, auto.
• title ( a caption).
• Similar to HTML links, but slightly more fancy.

70
XLink Simple links (2)

• Example 2
• anchor
• ltpoem_anchor xmllinksimple rolepoem312gtlttit
  legt Blue Mountainlt/titlegtlt/poem_anchorgt
• link
• ltrelated_poem xmllinksimple inlinefalse
  hrefpoem1 .xmlpoem312gtGoto related
  poemlt/related_poemgt
• Similar to HTML ltA NAMEpoem312gt.

71
Xlink Extended Link

• One link, several targets.
• For instance, the link See related poems would
  open as a list of links in a pop-up window.

72
Xlink Extended Link (2)

• Example
• ltrelated_poems xmllinkextended inlinefalse
  titleSee related poemsgt
• ltpoem_target xmllinklocator inlinefalse
  titleBlue Mountains hrefpoem1.xml/gt
• ltpoem_target xmllinklocator inlinefalse
  titlePink Flowers hrefpoem2.xml/gt
• ltpoem_target xmllinklocator inlinefalse
  titleSea of Green hrefpoem3.xml/gt
• lt/related_poemsgt

73
XPointer

• To define links that target points within
  documents.
• Special language to explain which spot is
  targeted.
• In HTML
• Need to manually insert a tag ltA NAMEgt.
• Hence need to own the document.
• With Xpointer
• No need to add anything to the target document.

74
XPointer (2)

• Example
• ltrelated_poem xmllinksimple inlinefalse
  hrefpoem1.xmlroot().child(2)gtGo to related
  poemlt/related_poemgt
• Other possibilities
• root().child(3).child(4)
• id(poem273)
• root().descendant(2, stanza)
• root().string(1, my heart)
• span(root().child(3), root().child(5))

75
Resource Description Framework (RDF)

• Defines syntax for describing resources.
• Metadata
• Similar to information in OPAC records.
• Essential for identification and retrieval of
  documents.
• Illimited nesting.

76
RDF Example

• ltrdfRDFgt
• ltrdfDescription about"http//www.w3.orggt
• ltPublishergtWorld Wide Web Consortiumlt/Publishergt
  
• ltTitlegtW3C Home Pagelt/Titlegt
• ltDategt1998-10-03T0227lt/Dategt
• lt/rdfDescriptiongt
• lt/rdfRDFgt
• Note Adapted from example in the RDF
  recommendation.

77
RDF Example (with nesting)

• ltrdfRDFgt
• ltrdfDescription about"http//www.w3.orggt
• ltPublishergtWorld Wide Web Consortiumlt/Publishergt
  
• ltTitlegtW3C Home Pagelt/Titlegt
• ltDategt1998-10-03T0227lt/Dategt
• ltCreatorgt
• ltPerson about"http//www.w3.org/staffId/85740"
  gt
• ltNamegtOra Lassilalt/Namegt
• ltEmailgtlassila_at_w3.orglt/Emailgt
• lt/Persongt
• lt/Creatorgt
• lt/rdfDescriptiongt
• lt/rdfRDFgt
• Note Adapted from example in the RDF
  recommendation.

78
RDF in practice

• RDF defines only the syntax, not the content.
• Can accommodate most document description
  schemes.
• Example of use Implementation of Dublin Core.
• Example of industry support ABC News, CNN and
  Time Inc.
• Both for HTML and XML documents.
• Details of implementation not entirely clear.

79
XML Schemas

• An alternative to DTDs.
• Still a working draft.
• Easier because uses the XML syntax.
• Data typing possible, unlike DTDs
• (integer, floating number, date, string, etc.).

80
XML Schema Example

• lt?xml version"1.0"?gt
• ltSchema name"poemSchema"
• xmlns"urnschemas-microsoft-comxml-data"
• xmlnsdt"urnschemas-microsoft-comdatatypes"
  gt
• ltElementType name"POEM" content"eltOnly"
  model"closed"gt
• ltelement type"TITLE"/gt
• ltelement type"AUTHOR"/gt
• ltelement type"STANZA" maxOccurs""/gt
• lt/ElementTypegt
• ltElementType name"TITLE" content"textOnly"
  model"closed" dttype"string"/gt
• ltElementType name"AUTHOR" content"eltOnly"
  model"closed"gt
• ltelement type"FIRSTNAME"/gt
• ltelement type"LASTNAME"/gt
• lt/ElementTypegt
• ...
• lt/Schemagt

81
Unicode

• Default character encoding for XML.
• Great improvement for encoding of non-western
  languages
• more than 65,000 characters,
• Eventually will represent all alphabets and
  writing systems,
• Also includes special typographic characters (
  ¼ ).

82
SGML, XML, HTMLWhat is the difference?

• XML SGML slightly simplified.
• HTML just an SGML DTD.
• Can be easily converted to an XML DTD.
• Relationship
• XML and SGML are meta-languages,
• HTML is a language.

83
Searching XML Documents
84
Models for XML RepositoriesFlat file system

• A bunch of XML documents in a folder.
• Native XML search engine
• an XML-aware Web site search engine.
• XML Query Language XQL
• Still in development
• Find word milk only when it appears in
  attribute DIETINFO2 of element PRODUCT.

85
Models for XML RepositoriesRegular relational
databases

• E.g., Web-based OPACs, Ovid, Amazon.
• Back-end relational DBMS
• MS Access or Oracle, for instance.
• Web interface
• uses scripts like CGI or Cold Fusion,
• Easy to change the scripts to output XML instead
  of HTML,
• Can even produce XML OR HTML according to the
  capabilities of the requesting browser.

86
Models for XML RepositoriesXML-aware relational
DBs

• Benefit from R-Databases AND XML advantages.
• Mixed record
• Nested structured text difficult to map to R-DB.
• However, many structured texts have a table-like
  section (the bibliographic information).
• R-Databases very mature technology (data
  integrity, security, load balance, etc.).

87
Models for XML RepositoriesXML-aware relational
DBs (2)

• Example of Oracle
• Enhanced full-text capabilities
• indexing,
• truncations, stemming, thesaurus, etc.,
• XML-like searching,
• can create SQL queries with embedded XML
  subqueries.
• Automatic mapping
• R-DB record --gt XML document,
• XML document --gt R-DB record,
• Virtual flat file system.

88
Information Retrieval Standardfor XML

• Needed to implement cross-repository search
• To query across several XML servers seamlessly,
• Whatever the implementation on the server side
  (Flat file system, R-DBMS, etc.).

89
Information Retrieval Standard Z39.50

• Used in the library community.
• To query OPACs, indexes, etc.
• Possible to specify
• A Query Language,
• The format of the results,
• A session protocol.

90
Information Retrieval Standard Z39.50 XML

• Currently beginning to integrate XML
• Defined as a possible output format,
• Some propositions to use XML as an alternative to
  BER for overall Z39.50 syntax.
• Once XQL is stabilized it could be ported to
  Z39.50.
• Good candidate to become the IR-Standard for XML.
• Little known outside the library community.

91
XML in Libraries
92
Which library projects are already using XML/SGML?

• Mostly academic institutions.
• (as well as Library of Congress and NYPL.)
• Usually in SGML.
• (Very recent ones in XML.)
• Mostly
• large and long-term digitization projects,
• involving the digitization of numerous texts.
• Converted to HTML on-the-fly.

93
Text Encoding Initiative (TEI)

• Standard to encode primary sources in the
  Humanities.
• SGML-based. (It is an SGML DTD.)
• Currently being converted to XML.
• Maintained by TEI Consortium.
• Widely adopted in Humanities computing community.
• Has spread to libraries.

94
Examples of TEI Projects

• Special collections
• Library of Congresss American Memory Project,
• Literary texts
• U. of Virginias E-text Collection,
• Browns Women Writers Project,
• Historical editions (MEP DTD)
• Abraham Lincoln Papers,
• Susan B. Anthony Papers.

95
Encoding Archival Description (EAD)

• Finding Aids to Special Collections and Archives.
• SGML/XML-based standard. (It is a DTD.)
• Maintained by the Library of Congress.
• Widely adopted.

96
Examples of EAD Projects

• Among many others California Digital Librarys
  Online Archive of California.
• Union DatabaseRLGs Archival Resources Project
• (MARC AMC records and EAD finding aids).

97
Materials Used by Libraries

• Reference Materials
• Oxford English Dictionary,
• American National Biography,
• Electronic Journals
• Springer-Verlags Link.

98
XML in Libraries What will it change? (1)

• EAD finding aids
• Offer precise and controlled search capabilities,
• Make the creation of union databases possible.

99
XML in Libraries What will it change? (2)

• Full-text databases of primary sources
• Easy to search, display, etc.
• Next step, union databases.
• With precise and controlled search capabilities.
• Full-text databases of e-journals, monographs.
• Competition with PDF/page images, though.
• Again next step, union databases.

100
XML in Libraries What will it change? (3)

• More sophisticated and customized clients
• Bibliography manager,
• Concordance program.
• New library standards based on XML
• TEI, EAD
• MARC (!)
• XML not just a fad, more than 10 years of
  SGML-based TEI.

101
XML in Libraries What will it change? (4)

• XML is more likely than any other formats to
  resist obsolescence
• Platform independent,
• Open standard
• (not proprietary),
• Written in ASCII/Unicode plain text
• (no binary encoding, the simplest text editor can
  read it),
• Tags are human-readable.

102
Web-based referenceWhat will XML change? (1)

• Topic-specific meta-search engines
• e.g., job search or book search.
• Will become ubiquitous.
• Already exist but awkward for developers.
• Small communities --gt can agree on a DTD.
• See agreement CNN all on RDF.
• Can also work without common DTD.
• In the vendors interest.
• All database-based --gt easy conversion.

103
Web-based referenceWhat will XML change? (1)

• General search will not improve for a long time
• A lot of legacy data
• The whole current WWW!.
• Numerous users will not switch to XML.
• Especially basic users.
• How to deal with thousands of different DTDs?

104
Should you use XML in your project today?

• Are your data made of a repetition of similar
  objects? (e.g., 3000 poems)
• Is your project database-based?
• Is your project large?
• Do you plan to
• deliver to different output devices?
• integrate your project with others? (e.g. union
  database)
• develop advanced capabilities? (server-side or
  client-side)