Meeting on Mansi fst
- nov 2024, present Csilla, Jack, Trond
Agenda
Status
Marginal increase in text coverage + drop in speller suggestion (20% down, probably some technical problem).
Done:
Csilla:
Has been working on missing list:
- 281 TBS (to be solved) + 407 lines of Unexplained_missing.txt.
Jack
Not work on mns lately
Trond
Has bee working on setting up divvunspeller, and succeeded (but with
Report to FM
Content:
- Present the good coverage
- Present the somewhat not that positive result of the suggestion mechanism
- Present the general benefit (possible to improve, to widen to new applications or new lgs)
Effect:
- if the speller is made known to the Mansi speakers and the format is well-known, it will have an impact on the whole speaker community
- for the journalists, the program will in any case be helpful
Work ahead
Russian loanwords
If it differs in any way from the Russian dictionary word, we should add it to the
диалогыт –≥ диалог is Mansi
When it is (a hapax) enclosed in quotation marks, it is Russian
More nouns
Še have 4000
Hoš to proceed:
1. reverse-sort missing list and add large chunks (half of them perhaps nouns)
2. Munkacsi dictionary: Write words we suspect are not in the fst but Csilla knows (are in use)
3. …
The TBS words
:%>%{аяØ%} PRFPRC ; ! атхатам XXX correct but unrecognised
+Err/Orth:%{VU%}%>%{уюØ%} PRFPRC_PLAIN ; ! adding ad lib
- Csilla to consult with Jack and/or Trond ad lib on TBS and other upcoming problems
- Jack to critically look at the V_UBA lexicon (the -aм participles, the lexicon overlaps, ..)
- Jack to look at the TBS
Suggestion mechanism
- Trond to investigate the drop, thereafter return to this meeting.
- Csilla to find some typos for the y2024 file (e.g. typos from the textbook).