Finite state and Constraint Grammar based analysers, proofing tools and other resources
Work on speller evaluation, finally understanding the tests and now both long vowels and capital letters (Эспоо) are now ok.
Went down on the missing lst, down to 400: - added missing stems to stem list - when bad analyses, added the missing ones
Been to conferences. One talk on Mansi and the societal situation, one on corpus building.
Discussed with Csilla on postpositions and ordinal numerals (these things pretty much in place).
Now back on a big issue: adding long and short forms to endings.
Results, see test diary.
The results are good enough for use. We should advice key users to test the online speller. Thereafter we may publish the speller in DM. Before that, we would still like to:
Thus: Agitprop whenever after today, but launch in the autumn.
Ordinals +A+Ord, cardinals +Num (?). Have looks at [10.32] Trond Trosterud lang-mns/src/fst/transcriptions/*.lexc ad lib.
We want a new version of the missing list. We now do: 1. Csilla corrects postpositions and does not touch the Luima_Seripos_2013-2017.missing.freq.240208 file 2. Meanwhile, Jack finishes the reported and thus open issues in this file 3. When there are no more open issues, we make a new missing.freq file, perhaps from a bigger corpus (cf. next issue).
Børre has adjusted the borders. We must evaluate and then recovert.
The column-final hyphens are still problematic, we must look into that with Børre.
To be decided upon, appr in two weeks.
News Flash!!
Do we want to do anything for this: IWCLUL 2024 at Metropolia, Helsinki? https://acl-sigur.github.io/iwclul2024.html
Probably yes. We return to this on next meeting.