Module talk:pra-decl/noun

@Wyang Hi, any idea why this isn't working? Most of it is copied from MOD:pi-decl/noun. —Aryaman^{A (मुझसे बात करें • योगदान)} 03:28, 20 January 2018 (UTC)

@AryamanA Made it output some more meaningful stuff on the doc page now. Wyang (talk) 03:46, 20 January 2018 (UTC)

@Wyang: Oh, thanks!! I guess I made some dumb mistakes lol. —Aryaman^{A (मुझसे बात करें • योगदान)} 16:08, 20 January 2018 (UTC)

Diaeresis Diacritic

@AryamanA If the transliteration of a word has a diaeresis diacritic, then the declension produces errors such as at 𑀕𑀅 (gaa) and 𑀮𑀆 (laā). Also, is pse a typo for psu? Kutchkutch (talk) 08:19, 19 July 2020 (UTC)

I'm temporarily displaying the declensions of 𑀕𑀅 (gaa) and 𑀮𑀆 (laā) at User:RichardW57/sandbox. I can't see any problem. Has it now been fixed? --RichardW57 (talk) 01:57, 14 May 2022 (UTC)

@Kutchkutch: I should have pinged you with the question above. --RichardW57m (talk) 13:51, 16 May 2022 (UTC)

Roadmap

@Kutchkutch, Svartava: What is the road map for Prakrit declension? As you are pivoting through the Roman script (with no support for idiosyncratic spellings), why not consolidate with one code module for all dialects and scripts and per-dialect data-only modules? I would make Module:pmh-decl-noun-knda and Module:pmh-decl-noun-deva redundant. This module (Module:pra-decl/noun) can do the script detection. It makes sense to have one template per language, possibly with others for handling irregularities in declension. --RichardW57m (talk) 13:51, 16 May 2022 (UTC)

Centralising the code will make automating testing easier. --RichardW57m (talk) 13:51, 16 May 2022 (UTC)

@RichardW57, RichardW57m: The only overview of Prakrit declension so far is from Category talk:Prakrit languages:

The data module for regular Magadhi declension is at: MOD:pra-decl/noun/inc-mgd

The data module for regular Maharastri declension is at: MOD:pra-decl/noun/pmh

Data modules for regular Sauraseni and Ardhamagadhi declension have not been created yet.

There is a template for romanised masculine a-stem Sauraseni declension at T:psu-decl-noun-a-m

with the addition of Template:pmh-decl-adj. Kutchkutch (talk) 10:24, 21 May 2022 (UTC)

@Kutchkutch, Svartava:: By road map, I meant plan for extending the coverage. There's a hint at User:Kutchkutch#Devanagari in Prakrit declensions, which seems to include the policy that any irregularities require that the entire declension be specified manually. I suppose that could be done by cut and paste from the automatically generated regular declension, e.g. to add the irregular accusative singular 𑀲𑀸𑀫𑀺𑀡𑀁 to 𑀲𑀸𑀫𑀺 (sāmi).--RichardW57m (talk) 11:21, 23 May 2022 (UTC)

What I propose to do is to:

1. Merge Devanagari capability into Module:pra-decl/noun.
2. Test it against Module:pmh-decl-noun-deva. I will keep the tests visible.
3. Once it works at least as well, replace the former by the latter in the templates.
4. Request the latter's deletion.

I intend to do the same for the Kannada script.

Silence is consent. --RichardW57m (talk) 11:21, 23 May 2022 (UTC)

@RichardW57, RichardW57m: There has never been a comprehensive plan for extending the coverage since many of the declension paradigms may not be understood well enough to add a large amount of data at once, especially for certain lects. This may be due to the fragmentary attestation in the primary sources and the unclear documentation in the secondary sources. Currently, irregular declension is specified manually for five entries using Template:pmh-decl-noun-irregular. Kutchkutch (talk) 00:36, 24 May 2022 (UTC)

@Kutchkutch, Svartava, AryamanA: This module is set up to handle multiple dialects by means of data files, which can be extended lect by lect and paradigm by paradigm. I notice that at present the Magadhi data file omits the feminine declensions. (They can even be extended case/number combination by case/number combination if appropriate.) All that is missing is script by script, though at the module level the coverage would be at the same for all scripts. Blocking combinations of script and lect could be done at the template level if, as at present, each lect had its own declension template. We could also add such blocking in the module itself, though it would be ugly. --RichardW57 (talk) 08:05, 24 May 2022 (UTC)

Actually, the blocking in the module needn't be so ugly. The code could look like:

if blocked[dialect] and blocked[dialect][sc_code] then error(...

with a table blocked set up as a sparse boolean array defaulting to nil. --RichardW57m (talk) 09:42, 24 May 2022 (UTC)

@Kutchkutch, Svartava, AryamanA: Well, I've repurposed {{psu-decl-noun}} to generate Sauraseni declensions. Feel free to review the data module Module:pra-decl/noun/psu; I've appended sample declensions to Module:pra-decl/noun/gallery/documentation. --RichardW57 (talk) 23:04, 4 June 2022 (UTC)

The above discussion has been rendered confusing by the renaming of the files in September 2023. There is now an up-to-date table. --RichardW57 (talk) 20:46, 19 September 2023 (UTC)