XML meeting, November 23rd, 2009

• Present: JB, RC, PD, JK, FJK, MS, WS, KT, JW, DW
• Protocol: KT
• Language technology
  • Greek
    • JW: customized two scripts for Greek to transcode from Betacode to Unicode
      • Macron problem still there
      • WS: Capital letters problem solved
  • Arabic
    • JW: first used Aramorph system
    • found an Extended Buckwalter on the net
      • WS: Extended Buckwalter is for a Qur'an project, not sure, if applicable
    • Hyphen problem
      • MS: hyphens are Buckwalter artefacts: should not be displayed (but not deleted either)
      • PD: hyphens come from filemaker file (used by transcriber to represent blanks)
    • JB: are there other characters to be taken care of?
      • JW: yes, works on it with MS
    • DW: Writing direction can be set in XHTML
• Normalizing
  • RC: please document normalizing steps
  • DW: make overview of the architecture
• Work on Liddell Scott Jones
  • WS: final sigma is wrong in some places
    • possible error source: new Lex script
  • Problems in transcoding should have been solved already by Perseus
• Priorities
  • 1. edo/ sum-problem
  • 2. Validation
  • 3. role of eXist repository
  • 4. new eXist version
  • 5. eSciDoc
  • 6. Transcoding issues
• DTD/ RNG
  • WS: implicit validation only through Schema and DTD
    • explicit validation through RNG in eXist using Jing
  • How to treat the DTD fragment
    • DTD fragment supports validation and saves 10% of the documents' size
    • Possible solution: separate validation and display of XML
    • DW: in general, entities are difficult in XML, it is better to resolve them
      • DTD fragment generates errors in a big number of tools
    • WS: resolving makes xml file harder to read and edit
    • to have xml file in eXist without fragment, a simple XSLT script is sufficient
    • WS: nice to have conversion back to version with fragment
      • can be done via script
• Disambiguation problem
  • Highest priority
    • Possible solution: introduce hyperlemma
• Pollux: change to new dictionaries?
  • JB: updates are not only corrected but have also different structure
    • dicitionaries would have to be transformed
  • DW: why do we not use the LSJ hosted at Perseus?
    • JW: we do not know if the corresponding entries are there
      • Perseus is slow
• Abbreviations
  • WS: two different kinds of abbreviations
    • book specific and commonly used ones
    • should offer service
    • PD: make enhancement of morphological server: offer also abbreviation resolver
      • similar to docspecs
    • schedule a meeting for that
      • WS collects material
    • PD: WS solves these problems for the third time
      • another task for the IT archaeological meeting
  • Also missing: MDHs article on linguistic middleware