140829

Present: Heli, Jaska, Trond

Status
Participants
Seuraava kokous

Program is online
650 venäjän sanaa
Burjaatin dokumentaatio ei vielä online (Trond)

Leksa

Lähtökohta:

650 sanaa vs. Kildin Saami list
70 Liivi, vuorimari, x
1791 Kiltinänsaame

Intersection of 1-3, or of 1+3 and 2+3.

main/ped/liv/src/A_liv2X.xml
main/ped/liv/src/N_liv2X.xml
main/ped/liv/src/V_liv2X.xml

Sanat + semanttinen luokittelu

TODO: (Heli)

Suunnitella Leksa-malli valmiiksi täytettäväksi.
Runko tehdään valmiiksi xml-muodossa sjd: n perusteella.

utgångspunkt:

sjd -> rus, eng, ...
- fjern sjd, fyll ut med eigne ord
erzya -> rus, fin, eng
sen jälkeen käännetään koneellisesti

sjd * intersection 1, 2 men fjern lemmat nya versioner => endre "sjd" till "xxx" etc.

   <e>
      <lg>
         <l pos="n">барбанн</l>
      </lg>
      <sources>
         <book name="l2"/>
      </sources>
      <mg>
         <semantics>
            <sem class="SCHOOL_EDUCATION"/>
         </semantics>
         <tg xml:lang="rus">
            <t stat="pref">барабан</t>
         </tg>
         <tg xml:lang="eng">
            <t stat="pref">drum</t>
         </tg>
         <tg xml:lang="fin">
            <t stat="pref">rumpu</t>
         </tg>
         <tg xml:lang="deu">
            <t stat="pref">Trommel</t>
         </tg>
      </mg>
   </e>

Languages that do not have Oahpa lexicon yet: bxr, izh, olo, mdf, mhr, mrj!, udm

something: kpv, myv, yrk

Undervisningsspråk är ryska, hjälp på andra språk.

Maks 15

bxr
- Jargal Badagarov, Ulan Ude (rus, eng)
izh
- Heinike Heinsoo, Timo Rantakaulio? (fin, rus, eng)
kpv
- Galina Punegova, StPb (fin, rus, eng) ?
- ?Galina Misharina?, Hki (fin, rus, )
- Svetlana Lumme (rus, fin, eng?)?
mdf
- Oleg Kazanin
mhr
- Sveta Hämäläinen, Syktyvkar (rus, eng?)
mrj
- Julia Kuprina, Patrick O'Rourke, Hki (rus, fin, eng?)
myv
- Ivan Ryabov, Saransk (rus, deu)
- Jelena Klementjeva, Saransk (rus)
olo
- Giloeva, Jsuu (fin, rus)
yrk
- Lotta Jalava, Hki (fin, eng, rus)
- Sven-Erik Soosaar (?Laptander)
udm
- Svetlana Yedigarova, Hki (fin, rus)
- Nadi Muš, Tln (est, rus, fin)

Usernames + passwords: Hki + GT

TODO (Heli) Restart: oahpa.no/erzya oahpa.no/yrkoahpa

Renew the FSTs: kpvoahpa/numra/cardinals /ordinals

main/langs/kpv/src/

TODO (Heli)

add link to oahpa.no/davvi on the front page of all alpha-oahpas
svn up for all oahpas before the course
compile the fsts and copy to /opt/smi/ before the course
Links to NDS dictionaries

4.9.2014 9: 30 Finsk tid

0000

Meeting, AKU Oahpa, 4.8.2014

Present: Heli, Jaska, Trond

Participants
Language of instruction
Budget
Course goals
Course planning day by day
Documentation
List of Languages
Next meeting

bxr
- Jargal Badagarov, Ulan Ude (rus, eng)
izh
- Timo Rantakaulio? (fin, rus, eng)
kpv
- Paula Kokkonen?, StPb (fin, rus, eng) ?
- Galina Misharina?, Hki (fin, rus, )
- Enye Lav? (rus, fin, eng?)?
mdf
- -
mhr
- Andrei Chemyshev?, Syktyvkar (rus, eng?)
mrj
- Julia Kuprina, Hki (rus, fin, eng?)
myv
- Ivan Ryabov, Saransk (rus, deu)
- Jelena Klementjeva, Saransk (rus)
olo
- Giloeva?, Jsuu (fin, rus)
yrk
- Lotta Jalava, Hki (fin, eng, rus)
udm
- Svetlana Yedigarova, Hki (fin, rus)
- Nadi Muš, Tln (est, rus, fin)
vep
- -

TODO:

Check the ones on the list and fill in evt. empty slots (Jaska)

Preferably Russian
Slides in Russian, talk in English at least in the beginning

Costs

Travel (Jaska)
Accommodation (Jaska)
Salary, Heli (Tromsø)

Income

AKU, UiT, own financing?

TODO: Specify costs (T, J) Look at financing (T, J)

How do we create linguistic content for Oahpa

During the week we will

set up Oahpas for the participating languages at the following level
- Numra (Evaluate, a preliminary version is set up on beforehand)
- Leksa (Words, semantic sets !!)
- Morfa-S (Noun case-number and verb setup)
- Morfa-C (Make some frames, set them up and see them work)
plan for further work on the respective Oahpa versions
plan how to integrate Oahpa in course curricula

Preparing before the course

Setting up user accounts, basic SVN.

Day 0:

Preliminary course in Unix, svn, etc, for people not having done this before.
All participants shall have checked out (at least) the ped catalogue and a working version of their own fst on their own machine (cf. the getting started page.)
basic SVN course. What is it, how to update, check in, etc.

Day 1:

Introduction

Giellatekno overview (infrastructure, projects, tools, Oahpa) (Heli, Trond)
Presentations by the participants about their languages and existing resources (textbooks a.o. teaching materials, corpora, language technology tools)

Leksa

Creating word lists in csv format.
Checking in new files and updating the existing ones.
svn ci -> Heli updates the db -> online check.
How to choose the vocabulary for Leksa. Textbook word lists, frequency dictionaries.

Day 2:

More Leksa
Morfa-S
- Drafting the case/number and person/number/tense forms to be included (Individuals should have this all thought out before hand ... )
- Evaluating the fsts
- Setting up the infrastructure for the respective exercises

Homework by Thursday:
- think about the productive Morfa-C frames in your language - which cases, which verbs?

Day 3:

hands-on

Numra

Evaluating existing Numras
How to improve them:
- Learning how to correct automata
- Learning how to extend Numra to ordinal, date, clock.

Leksa, Morfa-S

Continuing work

Day 4:

hands-on

Setting up Morfa-S. Case list, possible additional menus.
Writing Morfa-C frames.
Extending Leksa: Place names. The names that are different in the indigenous language

Day 5:

General discussion
- New ideas, thoughts that have come up while implementing Oahpa for your language.
- Discussing how to integrate Oahpa in language courses. Presenting kursa
Summing up and future work
- How to proceed with the development of your Oahpa.

TODO:

Work with the content of the program (orgkom)

PRELIMINARY READING etc. (tasks for participants)

Links on the Giellatekno pages
Look at an existing Oahpa, for Kildin Saami:
1. Kildin Saami Oahpa
2. The underlying source files
Look for and take along some learning materials (textbooks, workbooks, dictionaries). Best if they (also) exist in electronic format but paper format is also ok.
Think about the issues that need special pedagogical focus in your language, e.g. using some case(s).
Look at some Oahpa instances online (North Saami Oahpa, testing.oahpa.no/rusoahpa, testing.oahpa.no/fkv_oahpa, testing.oahpa.no/crk_oahpa), get inspiration and think about the analogies/differences with your language. (?)

The Oahpa pages for developers

... are today in English. In Russian as well? We want the "for developer" pages in Russian as well.

TODO:

Set up dummy files (Trond)
Machine translate + correct translation (barnraising project:all)
Integrate result in existing pages (links from above) Trond, Heli

Documentation pages for each Oahpa version

Set up dummy pages (Trond)

Languages with Oahpa setup

bxr - no setup
izh - testing.oahpa.no/izh_oahpa
kpv - oahpa.no/kpvoahpa
mdf - testing.oahpa.no/mdf_oahpa
mhr - testing.oahpa.no/mhr_oahpa
mrj - testing.oahpa.no/mrj_oahpa
myv - oahpa.no/erzya
olo - testing.oahpa.no/olo_oahpa
udm - testing.oahpa.no/udm_oahpa
vep - testing.oahpa.no/vep_oahpa
yrk - oahpa.no/yrkoahpa

TODO:

Set up bxr_oahpa (Heli)

Languages with transcriptors for Numra

bxr - no
izh - no
kpv - yes
mdf - yes
mhr - yes
mrj - yes
myv - yes
olo - yes
udm - no (numbers, clock Jaska)
vep - no
yrk - yes

TODO:

Make at least ordinals for the missing ones (Jaska, Trond) bjargal@mail.ru

Status for automata for the different languages

Grades:

A = comprehensive (speller quality)
B = Good (can be basis for text analysis, albeit with errors)
C = Basic vocabulary (expected to generate most Morfa-S words)
D = Parts of the vocabulary (Generates only part of what Morfa-S wants)

Status:

bxr - D
izh - C
kpv - B
mdf - B
mhr - B
mrj - B
myv - B
olo - C
udm - B
vep - D
yrk - C

Place names for leksa

bxr -
izh -
kpv -
mdf -
mhr -
mrj -
myv -
olo -
udm -
vep -
yrk -

Languages for which there are concrete plans for Oahpa work

bxr -
izh -
kpv -
mdf -
mhr -
mrj - yes (Kuprina Course)
myv - yes (Jerina course)
olo - yes (Giloeva course book)
udm -
vep -
yrk -

Aug 15th 0900 Swedish time.