Difference between revisions of "User:Firespeaker/TODO"
Jump to navigation
Jump to search
Firespeaker (talk | contribs) m (→Things for spectie: oops) |
Firespeaker (talk | contribs) (→Big) |
||
| (12 intermediate revisions by the same user not shown) | |||
| Line 1: | Line 1: | ||
See [[User:Firespeaker/TODO/done|TODO/done]] |
|||
== General == |
|||
* [[Apertium-kaz-kir/TODO]] |
|||
* [[Apertium-tat/TODO]] |
|||
* [[Apertium-kaz/todo]] |
|||
* [[Apertium Turkic/TODO]] |
|||
== Big == |
|||
* Implement productive causative in apertium-kaz |
|||
* Implement productive causative in apertium-tat |
|||
* Implement ifi.evid correctly in <s>apertium-kaz</s>, <s>apertium-kir</s>, apertium-kaa |
|||
* Change {{tag|px}} to {{tag|gen}}{{tag|attr}} and {{tag|gen}}{{tag|subst}} in <s>Kazakh, Kyrgyz,</s> all Turkic |
|||
* Change reflexive pronoun endings to px* forms in all Turkic (uncomment in Kyrgyz, etc.) |
|||
* Figure out [[User:Firespeaker/Kazakh_negatives|Kazakh and Kyrgyz negatives]] |
|||
== To think about == |
|||
* Problems with new build process |
|||
** How can we do single-category testvoc now? |
|||
** How can we make vanilla transducers (without MT-specific "wrong" POSes) |
|||
** How can we count trimmed stems? |
|||
== Things for selimcan == |
== Things for selimcan == |
||
* [[Kazakh and Tatar#Twol related stuff]] |
* [[Kazakh and Tatar#Twol related stuff]] |
||
* [[Apertium-tat/TODO]] |
|||
== Things for spectie == |
== Things for spectie == |
||
* Implement new case/postposition system at [[Morphology_of_Kyrgyz_language#All_cases_table]] |
* Implement new case/postposition system at [[Morphology_of_Kyrgyz_language#All_cases_table]] |
||
** From conversation in logs as Thu 12 Jul 2012 01:35:14 AM EDT |
|||
* <s>Document non-finite verb types at [[Turkic_lexicon#Non-finite_verbs]]</s> |
* <s>Document non-finite verb types at [[Turkic_lexicon#Non-finite_verbs]]</s> |
||
** Integrate Turkish non-finite forms into [[Morphology of Turkish]] and reference it from [[Turkic_lexicon#Non-finite_verbs]] |
** Integrate Turkish non-finite forms into [[Morphology of Turkish]] and reference it from [[Turkic_lexicon#Non-finite_verbs]] |
||
| Line 12: | Line 36: | ||
== Things for hector2 == |
== Things for hector2 == |
||
====New list 2012.09.21==== |
|||
* Vowel harmony: tests/автан.yaml. There is a strange change of %{Ӑ%} to и. --[[User:Hectoralos|Hèctor Alòs i Font]] 19:05, 21 September 2012 (UTC) |
* Vowel harmony: tests/автан.yaml. There is a strange change of %{Ӑ%} to и. --[[User:Hectoralos|Hèctor Alòs i Font]] 19:05, 21 September 2012 (UTC) |
||
:: Where is this strange change? The only thing not working for the автан.yaml file right now is a couple dative forms, but this is a widespread problem (see below). —[[User:Firespeaker|Firespeaker]] 03:29, 28 September 2012 (UTC) |
|||
=== questions/requests for hector2 === |
=== questions/requests for hector2 === |
||
* <s>What is special about [[пуртӑ]]? As far as I can tell, it behaves as one would expect for a noun ending in ӑ that has gemination. —[[User:Firespeaker|Firespeaker]] 07:56, 18 September 2012 (UTC)</s> |
|||
** The gemination of т. There is no gemination in CCӑ, only in VCӑ. The exception is пуртӑ.--[[User:Hectoralos|Hèctor Alòs i Font]] 09:14, 18 September 2012 (UTC) |
|||
::: Ah. I think we're going to move gemination to something conditioned by lexc. Trying to set up phonological triggers for it is too complex and has been causing too many problems. Marking nouns that get gemination in lexc is simple. —[[User:Firespeaker|Firespeaker]] 07:07, 19 September 2012 (UTC) |
|||
* Add forms and correct ones at [[Morphology_of_Chuvash#Nouns_ending_in_.E2.80.B9.D0.BE.E2.80.BA|Nouns ending in о]] —[[User:Firespeaker|Firespeaker]] 08:05, 18 September 2012 (UTC) |
|||
** I'll do. In fact, the rule is quite simple: words ending in unstressed ''o'', ''u'' or ''a'' behave as words ending in ''ӑ'', except in gemination (and the "dictionary" form, which conserves the Russian orthography). A subrule, which complicates this general one, is that the 3rd person differs for words invariable in Russian, and also радио seems to be special (it behaves as ''o'' was stressed - it seems there some Russian words are assimilated changing the stress to the final vowel). I'll write down the examples, change the wiki, etc.--[[User:Hectoralos|Hèctor Alòs i Font]] 19:59, 18 September 2012 (UTC) |
|||
*** Done. There are two columns I don't understand: ''σ́]'' and ''о→ов?'', so I'm not sure what to put there. I've added a new column, about the genitive form. I've changed some words and added new ones, in order to have all types of words (I think) and to use these words that are in the wiki. All words have their yaml file.--[[User:Hectoralos|Hèctor Alòs i Font]] 06:41, 22 September 2012 (UTC) |
|||
* Could we have pages for the following irregular nouns, especially the first two? —[[User:Firespeaker|Firespeaker]] 08:30, 18 September 2012 (UTC) |
* Could we have pages for the following irregular nouns, especially the first two? —[[User:Firespeaker|Firespeaker]] 08:30, 18 September 2012 (UTC) |
||
** [[пичче]] |
** [[пичче]] |
||
| Line 34: | Line 49: | ||
** [[ен]] |
** [[ен]] |
||
** [[хӗрри]] |
** [[хӗрри]] |
||
* <s>Could we have a page for the following regular noun? —[[User:Firespeaker|Firespeaker]] 08:48, 18 September 2012 (UTC)</s> |
|||
** <s>[[хӗв]]</s> |
|||
*** вот! There's also a yaml file. I am not sure yet about px2sg.loc and px2sg.abl, but all other forms should be correct. |
|||
**** Awesome, thanks! |
|||
***** In principle the forms px2sg are already revised. However I'm not really sure about px2sg.dat. The forms px2sg are seldom used and there were zillions of errors in px2sg.loc and px2sg.abl, even after 2-3 revisions (i.a. by philology teachers). In px2sg.dat in some cases {У} falls, but I'm not sure which is the rule (I haven't analyzed it yet) and I guess the informants may have made mistakes. In these cases, rarely used, where a few cases without any difference with others don't follow the rules, I finally decided to follow the general rule and not believe the informants.--[[User:Hectoralos|Hèctor Alòs i Font]] 19:03, 20 September 2012 (UTC) |
|||
* Could I have yaml files for the following words? —[[User:Firespeaker|Firespeaker]] 08:46, 18 September 2012 (UTC) |
* Could I have yaml files for the following words? —[[User:Firespeaker|Firespeaker]] 08:46, 18 September 2012 (UTC) |
||
** <s>[[йӗп]]</s> |
** <s>[[йӗп]]</s> |
||
| Line 48: | Line 58: | ||
*** You can use the script tests/gen_yaml_mot.sh, which creates a yaml for a given noun from the wiki.yaml file (which I regularly read from the wiki, when I change something in it). As above, I still have some doubts about px2sg.loc and px2sg.abl, but all other forms should be correct. Атте should be correct in all forms.--[[User:Hectoralos|Hèctor Alòs i Font]] 19:07, 18 September 2012 (UTC) |
*** You can use the script tests/gen_yaml_mot.sh, which creates a yaml for a given noun from the wiki.yaml file (which I regularly read from the wiki, when I change something in it). As above, I still have some doubts about px2sg.loc and px2sg.abl, but all other forms should be correct. Атте should be correct in all forms.--[[User:Hectoralos|Hèctor Alòs i Font]] 19:07, 18 September 2012 (UTC) |
||
**** See above about px2sg (by the way, I see that both атте and анне have the forms px2sg.loc and px2sg.abl which I corrected everywhere. As they are irregular words I don't dare correct them... It may be said that really in these words personal suffixes are used, so this gives some more confidence).--[[User:Hectoralos|Hèctor Alòs i Font]] 19:03, 20 September 2012 (UTC) |
**** See above about px2sg (by the way, I see that both атте and анне have the forms px2sg.loc and px2sg.abl which I corrected everywhere. As they are irregular words I don't dare correct them... It may be said that really in these words personal suffixes are used, so this gives some more confidence).--[[User:Hectoralos|Hèctor Alòs i Font]] 19:03, 20 September 2012 (UTC) |
||
* |
* I don't understand what the rule is for when {У} deletes in {{tag|px2sg}}{{tag|dat}}. Do you know what the rule is? —[[User:Firespeaker|Firespeaker]] 03:29, 28 September 2012 (UTC) |
||
: <pre>утӑ:ут%{nodup%}ӑ N1 ; ! "сено" ! exception according to И.П. Павлов 1974: 18</pre> |
|||
** I'm sorry. I didn't notice it was "especial". I've corrected the third person forms, which don't have to regular gemination. |
|||
* As you requested, I've put the files tests/v_кил_del.yaml and tests/v_кала_del.yaml where you can find examples of %{Ӑdel%}, %{ӗdel%}, %{иdel%}. кала is not working because of the vowel harmony problem you can find e.g. in tests/автан.yaml (my first request of the new list). --[[User:Hectoralos|Hèctor Alòs i Font]] 20:05, 21 September 2012 (UTC) |
|||
== General cv.twol TODO list == |
== General cv.twol TODO list == |
||
| Line 60: | Line 67: | ||
* "Irregular" nouns (family relations that have different stems):w |
* "Irregular" nouns (family relations that have different stems):w |
||
* clean up twol conflicts |
* clean up twol conflicts |
||
[[Category:TODO lists]] |
|||
Latest revision as of 21:32, 19 August 2015
See TODO/done
Contents
General[edit]
Big[edit]
- Implement productive causative in apertium-kaz
- Implement productive causative in apertium-tat
- Implement ifi.evid correctly in
apertium-kaz,apertium-kir, apertium-kaa - Change
<px>to<gen><attr>and<gen><subst>inKazakh, Kyrgyz,all Turkic - Change reflexive pronoun endings to px* forms in all Turkic (uncomment in Kyrgyz, etc.)
- Figure out Kazakh and Kyrgyz negatives
To think about[edit]
- Problems with new build process
- How can we do single-category testvoc now?
- How can we make vanilla transducers (without MT-specific "wrong" POSes)
- How can we count trimmed stems?
Things for selimcan[edit]
Things for spectie[edit]
- Implement new case/postposition system at Morphology_of_Kyrgyz_language#All_cases_table
- From conversation in logs as Thu 12 Jul 2012 01:35:14 AM EDT
Document non-finite verb types at Turkic_lexicon#Non-finite_verbs- Integrate Turkish non-finite forms into Morphology of Turkish and reference it from Turkic_lexicon#Non-finite_verbs
Get numerals working in KazakhAdjectives in Turkic lexicon.
- Good enough? —Firespeaker 15:52, 21 September 2012 (UTC)
- Could we get examples/categories for kir/kaz ? - Francis Tyers 17:39, 21 September 2012 (UTC)
- Good enough? —Firespeaker 15:52, 21 September 2012 (UTC)
Things for hector2[edit]
- Vowel harmony: tests/автан.yaml. There is a strange change of %{Ӑ%} to и. --Hèctor Alòs i Font 19:05, 21 September 2012 (UTC)
- Where is this strange change? The only thing not working for the автан.yaml file right now is a couple dative forms, but this is a widespread problem (see below). —Firespeaker 03:29, 28 September 2012 (UTC)
questions/requests for hector2[edit]
- Could we have pages for the following irregular nouns, especially the first two? —Firespeaker 08:30, 18 September 2012 (UTC)
- Could I have yaml files for the following words? —Firespeaker 08:46, 18 September 2012 (UTC)
йӗппуртӑатте- пӳ
- лаша
утӑ- You can use the script tests/gen_yaml_mot.sh, which creates a yaml for a given noun from the wiki.yaml file (which I regularly read from the wiki, when I change something in it). As above, I still have some doubts about px2sg.loc and px2sg.abl, but all other forms should be correct. Атте should be correct in all forms.--Hèctor Alòs i Font 19:07, 18 September 2012 (UTC)
- See above about px2sg (by the way, I see that both атте and анне have the forms px2sg.loc and px2sg.abl which I corrected everywhere. As they are irregular words I don't dare correct them... It may be said that really in these words personal suffixes are used, so this gives some more confidence).--Hèctor Alòs i Font 19:03, 20 September 2012 (UTC)
- You can use the script tests/gen_yaml_mot.sh, which creates a yaml for a given noun from the wiki.yaml file (which I regularly read from the wiki, when I change something in it). As above, I still have some doubts about px2sg.loc and px2sg.abl, but all other forms should be correct. Атте should be correct in all forms.--Hèctor Alòs i Font 19:07, 18 September 2012 (UTC)
- I don't understand what the rule is for when {У} deletes in
<px2sg><dat>. Do you know what the rule is? —Firespeaker 03:29, 28 September 2012 (UTC)
General cv.twol TODO list[edit]
geminationӳ:ӗв, у:ӑв- Nouns ending in о
<px2sg><dat>of nouns- "Irregular" nouns (family relations that have different stems):w
- clean up twol conflicts
