SourcesEdit

Hi Jim, you don't need to create an item for every issue of a journal. I've corrected Apertium: a free/open-source platform for rule-based machine translation (Q14508416). See: Help:Sources. Cheers --Kolja21 (talk) 11:55, 5 August 2014 (UTC)

GlobalFactSyncEdit

Thought: We should try to resubmit GlobalFactSync for support this time around. Sj (talk) 19:51, 14 November 2018 (UTC)

Success! If you are still interested :) Sj (talk) 21:53, 13 May 2019 (UTC)

Features for mutation on Irish LexemesEdit

Hi,

I see that you created definite article mutation (Q66587485) that you used as grammatical feature on uisce (L2093) and Polainn (L42281).

I'm wondering if this is the best way. For other Celtic Languages, we choose to use general items according to the "morphology" and not based on "grammar" (to say it simply). For example in Breton and Cornish, we didn't use "definite article mutation" (by analogy, the cause) but "soft mutation" (soft mutation (Q56648699), the effect). What do you think? Would it make sense for Irish too? (it would be nice to have a consistent model for similar phenomenon in Celtic languages ;) ).

See also this discussion Wikidata talk:Lexicographical data/Archive/2020/01.

Cheers, VIGNERON (talk) 11:57, 8 November 2020 (UTC)

@VIGNERON: It's not the same as soft mutation (that would be equivalent to lenition). It's specifically a "grammar" based property, the main difference being one of convention: Irish nouns in Wiktionary are also listed with their definite article forms, while (e.g.) Welsh nouns are not, though they do display similar variation (y gath for singular, y cathod for plural). -- Jimregan (talk) 14:58, 8 November 2020 (UTC)
Also, I should have additionally added the eclipsed forms as eclipsed forms; I think I stopped because I thought there was a reason why the properties needed for Irish were missing. -- Jimregan (talk) 15:43, 8 November 2020 (UTC)
Mhhh, yes. But what I was saying is not to use « "grammar" based property » for grammatical feature but "morphological" one for efficiency (like for cath (L21637) to take your cat example). But maybe it's not a good idea for Irish (Gaelic languages don't work exactly like Brittonic languages), I'll trust your judgment. Cheers, VIGNERON (talk) 20:16, 8 November 2020 (UTC)
Yikes! Don't trust my judgement - I certainly don't! I've been in and around the area of technology for Irish for a few years - long enough to have forgotten that The Right Thing To Do was not at all obvious to me, and that my first instincts were more or less the same as yours.
It would actually end up being less efficient. It's not as much of an issue in Welsh because (IIRC) you only need to know if the noun is feminine singular to mutate after the definite article, and don't otherwise, but in Irish you need to know the gender, case, number, first letter of the word (or, if it's 's', the letters that follow it). Dictionaries that give full paradigms give both forms with and without the article because of this variability, so it would be expected to have that information, if Wiktionary were to switch templates to use Wikidata. -- Jimregan (talk) 23:13, 9 November 2020 (UTC)
Ok, thanks for the answer. Indeed this item seems needed, I'll leave it like that (at last for the moment ;) ). Cheers, VIGNERON (talk) 10:52, 10 November 2020 (UTC)