Meeting with Polderland 26.6.2007


  • Peter Beinema
  • Sjur Moshagen


  • Since last time
  • Possible issues
  • Next meeting
  • Todo items for the next meeting

Since last time


  • currently looking into bug with test system with/without user dictionary
    • still not completed yet
  • Adobe: developing for CS3 - work continues.
    • hyphenator for Afrikaans/Windows running, working on Sami extensions
  • evaluating price for InDesign speller
  • bug 402: tokenization issue, buried deep in code, fix available before final version
  • bug 419 (compounding: 2 elements -> 5): will be solved this week
  • bug 447 (installation on terminal server): looks like configuration issue
  • working on self-starting windows installer
  • have a set of penalty setting rules (to be included in phonetic rules) ready to be sent
  • have had some discussions internally on SáMI -> Sámi sugg mechanism. Currently, we do not suggest all uppercase in any of our spellers
  • speller tester (command line speller) tokenization: no trivial change. Has to do with the fact that MS word also does some tokenization. - punctuation that can, or can not, be part of a token (.'-)


  • CS3 seems to be installed, or is on the way to being installed in at least the biggest printing houses, but I still need to check some of the other publishers
    • still one more to check, but it seems CS3 will be a safe lower end supported
    • also needs to get confirmation from the board
  • installed new version of mklex - waiting for the first compilation run with it to finish
  • our speller test bench is starting to become really good: ) - have a look at this page where we measure performance on different editing distances, suggestion rating, number of failed suggestions, etc.


  • suggestions for generated compounds
    • send examples
    • penalty levels can be set at the begining of the phonetic rules file - Peter will send some examples and documentation
      • on its way
  • compounding limit - now 2 elements, should be higher
    • Peter will change it to 5 - otherwise identical to the newest speller.
      • will be sent this week
  • windows installer should autostart
    • see bug report - being worked on
  • case correction:
    • MáRKANSLUSKA => MÁRKANSLUSKA, suggests: Márkanšiljuska and other words with only initial capital.
    • SáMI => SÁMI, suggests: Sámi
      • Peter will look into possible settings to adjust this
      • it has been discussed now, it will probably be fixed, but not immediately
  • full stop at end of word - command line speller
    • will check this and get back wrt tokenization diffs with Word
      • needs some more investigation
  • hyphenation lexicon update - the dic file
    • the dic file is exactly the same as the speller lex file
    • the pattern file is a different story, and can not be updated by Divvun. But there shouldn't be any need for updating it - the resulting patterns would be the same

Next meeting

Next week (3.7.) at the usual time.


  • check inconsistent speller behaviour depending on the existence of userdict or not PLD
    • Peter has started to look at this
  • PLD continue to create hyphenator for Adobe InDesign CS3
  • PLD try to find proper compiler version for Adobe Indesign CS2. Looks bad.
    • old version does not work, user group + Adobe confirms it no longer sold
    • check with Sámi publishing houses whether support for CS2 is still needed ( Divvun)
  • speller in InDesign/InCopy option (PLD)
  • see bugs 402, 419, 447 in the Divvun Bugzilla (partially PLD)
  • add capital suggestion bug to Bugzilla (Divvun)