On 7 June 2012 21:20, Bernard Chardonneau <[email protected]> wrote: > >> Date: Thu, 7 Jun 2012 02:15:38 +0100 >> From: "Jimmy O'Regan" <[email protected]> >> To: [email protected] >> Reply-To: [email protected] >> Subject: Re: [Apertium-stuff] ordinales in French for fr-es pair >> On 6 June 2012 22:39, Bernard Chardonneau <[email protected]> wrote: >> > The old version was not working : >> > >> > With old version that can be seen on the website : >> > >> > echo "D'abord le 1er, puis le 2ème, le 3ème et les suivants parmi >> lequel le 21ème" | apertium fr-es >> > Primeramente el @1er, después el 2*ème, el 3*ème y los siguientes >> entre el cual el 21*ème >> > >> > Not surprised of the result, the French paradigm did not have a logical >> > content. >> > >> >> It had nothing to do with the content of the French dictionary. I take >> 'did not have a logical content' to mean 'I did not understand the >> regexes'. In future, you should take that as a sign to proceed with >> caution. > > Not only. If Every ordinal number in French starting from 2 can be > built by adding "e" ou "ème" to the last digit (and IT IS like that), > it's not worth distinguishing multi digit numbrers finishing by 1, 2 > or 3. And in the paradigm you put again : > > - it means 11, 21, 1234561 etc... finish like 1(st) which is not true > in French. > - and I don't see in it you to recognise a number finishing by a digit > from 4 to 9 (or also finishing with 0). >
"I" don't do anything of the sort; I neither wrote the original regex, nor have any sort of emotional attachment to it. That set of regexes matches, albeit incorrectly, what's in the bilingual dictionary; the eo-fr regexes do not, and introduce generation errors in the es->fr direction. >> I fixed the errors, and included the alternative forms you added to >> the dictionary, but kept the original entries as the defaults (as they >> are much more frequent - e.g., 'le 1r' gets around 50,000 google hits, >> while 'le 1er' gets around 50,000,000), and kept the regexes, because >> they catch more. >> > > For 1er or 1r, I don't remember what was taugth to me at the primary > scholl. I recently found a webpage telling 1r was the best (which is > the choice in eo-fr pair), but I don't find it again. > > Anyway, the first problem is to accept the most usual syntaxes. For > 1er, no problem, I also write it like this. > > But for the other numbers, even if websites don't aggree with the > best choice : > http://www.languageguide.org/french/ordinal-numbers/ > http://www.langue-fr.net/spip.php?article239 > Thanks, that's useful. See current output, below. Frequency favours -e over -ème, as does Wikipedia. For 2, highest to lowest is 2e, 2ème, 2d/2de, 2nd/2nde (but the margins are not so wide). > and the apertium-es-fr.fr paradigm is very far from the combinations > used in French language. > > echo "D'abord le 1er, puis le 2ème qu'on écrit aussi 2nd, le 3ème, le 4ème et > les suivants parmi lequel le 11ème le 21ème et le 31ème." | apertium fr-es | > tee /dev/tty | apertium es-fr > Primeramente el 1.º, después el 2*ème que se escribe también #2.º, el 3.º, el > 4*ème y los siguientes entre el cual el 11*ème el 21*ème y el 31*ème. > Tout d'abord le 1er, après le 2**ème que s'écrit aussi ##2d, le 3e, le 4**ème > et les suivants entre lequel le 11**ème le 21**ème et le 31**ème. > Current output: Primeramente el 1.º, después el 2.º que se escribe también 2.º, el 3.º, el 4.º y los siguientes entre el cual el 11.º lo 21.º y el 31.º. Tout d'abord le 1er, après le 2e qui s'écrit aussi 2e, le 3e, le 4e et les suivants entre lequel le 11e le 21e et le 31e. > The same from French to English using Esperanto as a pivot language : > > echo "D'abord le 1er, puis le 2ème qu'on écrit aussi 2nd, le 3ème, le 4ème et > les suivants parmi lequel le 11ème le 21ème et le 31ème." | apertium fr-eo | > tee /dev/tty | apertium eo-en > Unue la 1a, poste la 2a ke oni skribas ankaŭ 2an, la 3a, la 4a kaj la sekvaj > inter kiu la 11a la 21a kaj la 31a. > First the 1st, later the 2nd that one writes also 2nd, the 3rd, the 4th and > the consequent between that the 11st the 21st and the 31st. > '11th' - French isn't the only language with misgenerated ordinals :) > Looking at your changes, I saw ordinal are processed in the bidix > as other "words". Before, I did not find that in the bidix. > Everything is a "word", even punctuation. -- <Sefam> Are any of the mentors around? <jimregan> yes, they're the ones trolling you ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Apertium-stuff mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/apertium-stuff
