Experiences of Document Transformations with XSLT and DOM

1
Experiences of Document Transformations with XSLT
and DOM

• Anne Honkaranta, Virpi Lyytikäinen, Pasi
  Tiitinen, University of Jyväskylä, Finland
• inSGML project

2
Content

• Poem Publishers, Inc.
• Poems
• Publishing environment
• Transformations
• Tranformation techniques
• Transformations in server-client environment
• Tranformations in Poem Publishers, Inc
• Challenges encountered
• Lessons learned

3
Poem Publishers, Inc.

• Fictional company
• Publishes Finnish poems on WWW
• Poems are authored in XML format according to a
  DTD
• The company offers the poets an authoring
  environment if so desired
• The poems can form collections

4
Poem.dtd
5
Publishing environment

• Microsoft IIS server v. 5.0
• Jscript, VBScript
• ASP 3.0
• DOM II
• Internet Explorer 5.5 or newer
• CSS Level 2
• MSXML 3.0

6
Transformation

• Changing/converting document
• format
• structure /information schema
• content organization
• filtering the content
• all the above
• Conversion, filtering, and transformation are
  sometimes used as synonyms

7
Why you need transformations?

• Authors need content-oriented DTD
• Different end-user devices
• When managing documents we need to have them in
  an optimal format for processing
• --gt three-step publication process
• authoring -- processing -- output

8
Transformation techniques
9
Transformations in client-server environment
(XSLT/DOM)

• Alternatives
• using PI in XML source document (c)
• (can be written to the source document on a
  web server)
• DOM-interface and DOM objects for loading the
  source XML and XSLT (c/s)
• using DOM-interface scripting language
  (Vbscript, Jscript) or Java

10
Transformation chain (an example)
Output HTML/ XHTML doc rendered by CSS
11
Exampleusing PI in source XML
12
Example using DOM-objectsXSLT
13
Example using VbscriptDOM
14
Transformation types tested in Poem Publishers,
Inc.

• XML-to-XML
• XML-to-HTML
• XML-to-XHTML

15
Transformation needs tested in Poem Publishers,
Inc.

• Tasks tested
• combining multiple source documents into output
  view (poemheader/footer, poem list, poem
  metadata)
• combining multiple source documents into one file
  (making a poem collection)
• combining XSLT transformation documents for
  transformation needs (poemfooter)

16
Example combining XSLT-stylesheets
lt?xml version1.0 encodingiso-8859-1?gt ltxsls
tylesheet xmlnsxsl"http//www.w3.org/1999/XSL/T
ransform" version"1.0"
xmlnsxlink"http//www.w3.org/1999/xlink"
xmlns"http//www.w3.org/1999/REC-html40
1"gt ltxslimport href"header.xsl"/gt ltxsloutput
method"html" encoding"ISO-8859-1" /gt
lt?xml version1.0 encodingiso-8859-1?gt !--
Filename header.xsl
--gt ltxslstylesheet xmlnsxsl "http//www.w3.org
/1999/XSL/Transform" version"1.0"
xmlnsxlink"http//www.w3.org/1999/xlink"
xmlns"http//www.w3.org/1999/REC-html4
01"gt ltxsloutput method"html" encoding"ISO-8859-
1" /gt ltxsltemplate match"/" name"header"gt
17
Challenges Encountered

• Problems with
• parsers and versions
• character encodings
• figures and links
• too many tools, scripting languages, and
  programs

18
Example Character encodings and parser

• MSXML

OUTPUT DOC
MSXML 3.0
INPUT DOC
-input doc encoding -maybe character
entities -entities are changed to actual
character reps. when transformed
-uses UTF-16 -detects output encoding from PI
when appropriate load/save methods
used -otherwise outputs UTF-16
-has some encoding -has an encoding
declaration -problem either of them is wrong
19
Possibilities

• you can use XSLT-stylesheets as components and
  combine them
• a stylesheet can be seen as a re-usable component
  on the server
• you can also chain transformations
• you can keep your data in content-oriented form
  and provide multiple output versions by using
  transformations
• problem management of DTDs, transformation
  components and versions

20
Lessons learned

• Use same character encodings in source documents
  and transformation scripts
• Offer a content oriented DTD for your authors
  there is propably need for transformations anyway
• Support level of CSS, XSLT and XML varies in
  browsers
• Tools are available for building XML publishing
  environments allow extra time for dealing with
  possible problems
• Multiple skills and tools needed in publishing
  environment, XML is not enough!

21
More information inSGML project
http//haades.it.jyu.fi/inSGML/