150430

Megan, Lene, Antti

1. Meeting with the teacher
2. itwêwina
3. Implementation on the gtoahpa server
4. Improvements
5. FST-issues
6. Next meeting

Meeting with the teacher

Megan has talked with Dorothy and here are some key words from the discussions:

• dealing with TA/TI forms (inverse forms), how to do that
• more instructions are needed
• MorfaC: some weird centences, e.g. plural mothers and fathers
• Numra: "mina" was missing in Money
• Leksa: more semantic fields in the menu
• NA and NI instead of AN and IN
• s/he as stat="pref"
• itwêwina is good

The solution of these questions comes later in this document.

itwêwina

itwêwina accepts and generates with hyphens and without hyphens

• niki-nohte-kakwe-nipan
• nikinohtekakwenipan

PV/nohte+PV/kakwe+nipâw+V+AI+Ind+Prt+1Sg	nikînôhtêkakwênipân
PV/nohte+PV/kakwe+nipâw+V+AI+Ind+Prt+1Sg	nikî-nôhtê-kakwê-nipân

conjugation: transitive inaminate 1s	ninitonên	ê-nitonamân
                                                    ênitonamân

Todo:

• Antti: check with Ryan treatment of preverbs in translation search

Implementation on the gtoahpa server

Possibly creating demo directly on the proper server? Many things needed to be redone.

Open issues - much TODO:

• audiofiles now missing from the section selection (Lene will talk to Ryan)
• New French files will be generated from original src files (no hurry)
• hints are now working, but will need scrutiny and further development based on user experience.
  • Lene will do some improvements based on discussions in the meeting
• instructions can be improved in the files in
  • crk_oahpa/crk_drill/templates/ (You can discuss with Ryan if you need help)
• NA/NI and s/he (Megan)
• Lene will improve Numra, but Megan sends an email with details
• semantic fields (Megan)
  • meta/semantic_sets.xml
  • crk_oahpa/crk_drill/forms.py

SEMTYPE_CHOICES = (
    ('HUMAN', _('human')),
    ('KIN', _('kinship')), <=== from src/N_crk.xml
    ('MIND-BODY', _('body and mind')),
    ('FOOD-DRINK', _('food and drink')),
    ('NATURE-ANIMALS', _('nature and animals')),
    ('THINGS', _('things')),
    ('DAILY-LIFE', _('at home')),
    ('OCCUPATION', _('occupation')),
    ('EVERYDAY', _('everyday')),
    ('DESCRIPTIONS', _('descriptions')),
    ('SPEECH', _('speech')),
    ('all', _('all')),
)

Megan will restrict the selection of nouns to MorfaC tasks by adding more semantic classes:

• PEOPLE, PLACE -- leksa
• mPEOPLE, mPLACE -- morfa - include for MorfaC

details: mîscakâs (only Pl)

Morfa-C: Px is restricted to only singular

 <q id="npx1">
    <qtype>N-PX</qtype>
    <question>
      <text>SUBJ cî ana</text>
      <element id="SUBJ">
            <grammar tag="N+AN+Sg+Px2Sg"/>
        <sem class="KIN"/>
      </element>
    </question>
    <answer> 
      <text>ehâ, N-PX awa</text>
      <element game='morfa' id="N-PX" content="SUBJ" task="yes">
        <grammar tag="N+AN+Sg+Px1Sg"/>
      </element>
    </answer>
  </q>

MORFA-C: Maybe only singulars?

 
   <q id="npl1">
    <qtype>N-PL</qtype>
    <question>
      <text>awîniki êhayâcik N-LOC</text>
      <element id="N-LOC">
        <grammar tag="N+IN+Loc"/>
        <sem class="PLACE"/>
      </element>
    </question>
    <answer> 
      <text>SUBJ N-LOC êhayâcik</text>
      <element game='morfa' id="SUBJ" task="yes">
            <grammar tag="N+AN+Pl"/>
        <sem class="mPEOPLE-PL"/>
      </element>
    </answer>
  </q>

gen_only="Sg"
<l pos="N" animacy="AN" rime="0" gen_only="Pl,Sg" audio="kohkos">kohkôs</l>
    <l pos="N" animacy="IN" rime="0" gen_only="Sg">kohkôsiwiyin</l>
    <l pos="N" animacy="AN" rime="0" gen_only="Pl,Sg,Loc">pahkwêsikan</l>

"rime"-codes are trigger codes!

Improvements for verbs

Verbs in Morfa S:
Menu: choose AI and TI (later also TA)

• Ex: to eat
  • AI mîcisow (eat)
  • TI mîciw (eat) kohkôsiwiyin (bacon) mîcimâpoy (soup)
  • TA mowêw (eat) oskâtâsk (carrot) kinosêw (fish) picikwâs (apple)
• Ex: to see
  • TI: wâpahtam (see) Sg3 mîcimâpoy
  • TA: wâpamêw (see) Sg3 kinosêw

Todo:

• Antti will add object to the verbs, like this:

<l pos="V" object="mîcimâpoy" trans_anim="TI">mîciw</l>

FST-issues

documentation

• -i- disappears in some diminutives (Lene)
• continue the work with FST - discussion
• conflicts can make the result of hfst and xfst different
• Add discussions and documentation of open questions to the documentation

Next meeting

• 4-8 recordings in Maskwacis
• 10-13 recordings in Maskwacis
• 14th Whit-Thursday