Samest meeting 21.11.2016

Participants: Fran, Heiki, Heli, Jaak, Jack, Kadri, Kaili, Trond


  • Estonian FST
  • Estonian Oahpa
  • MT
  • Võru Oahpa and FST
  • Papers/articles
  • Next meeting

Estonian FST


Not much has happened.

The exp-est.fst is now slightly more beautiful.


Work on generation done. Overgeneration has been a problem, is not that to the same extent.


Differences in coverage between langs/est and experimental-langs/est? Jaak is working on setting up a testing environment.

Estonian Oahpa

Heli has made the first attempt to generate the Oahpa database with Heiki's FST (experiment-langs).

  • The overall impression is good.
  • About 30 compound nouns and 50 simple nouns (those with Sg Nom homonyms) were not generated.
  • A few errors have been discovered.
  • Some specific discussions about preferred vs non-preferred forms with Heiki will follow.

The work on Morfa-C question-answer templates is in progress (a student from Tartu is working on it).


Both est-fin and fin-est now on http://gtweb.uit.no/jorgal/index.cat.html?dir=est-fin

URL: https://svn.code.sf.net/p/apertium/svn/incubator/apertium-fin-est

Summing up MT group meeting

We went through and fixed errors for finest

We also corrected spellrelax (both for est.fst and experimental-est.fst)

Online tools

Online at http://gtweb.uit.no/jorgal, but still not on http://gtweb.uit.no/mt/testing (perhaps not needed)





Issues with the Finnish fst:

minoriteettii: before the meeting it was analysed as min#ori#tee#tii

How to work with compound blocking:

+CmpNP/None = does not participate in compounding

example in the lexicon:

gaskka+CmpNP/Last+Sem/Dummytag:gas'kam8 SEAMU ;

Võru Oahpa and FST

Too much generation.



  • MT general paper (fmt, tt)
  • MT special aspects? (fmt, tt)
    • Assimilation/Dissemination?
    • Specific fin-est syntactic differences
  • MT system(s) --- sme-fin (fmt, tt)
  • Oahpa presentation (Heli)
  • Võru FST presentation
  • Estonian FST this-or-that way (Heiki)
  • FST and compounding for Germanic-like Uralic languages... (Heiki, tt?)
  • Est-Nob dictionary
  • Contrastive Linguistics (fmt, tt, Heiki ?)


  • NODALIDA papers due 20th Jan., demo papers 20th March
  • EAMT ???
  • ACL 2017 6th Feb

Next meeting

... in Helsinki, for the ones there. Wednesday, 7th December 13:00 Norwegian time

fmt has flight at 17: 05