> Date: Sat, 29 Dec 2012 22:56:33 +0100 > From: Jacob Nordfalk <[email protected]> > To: [email protected] > Reply-To: [email protected] > Subject: Re: [Apertium-stuff] Esperanto-bridged translation for distributed > social networks, especially for activists. > > 2012/12/29 FreeLab - Laboratorium Wolnosci <[email protected]> > > > > > As for bridge vs. direct translation, I think that we should at least > > start from the bridged model. It is more elegant (in terms of > > structure) and needs less resources. > > > No. > Double translation will double CPU and RAM resources needed and degrade > quality. > And, installing all pairs just requires <100MB disk space anyway. > >
I understand the idea of Freelab, and in fact, I had the same. I don't know what website is interested by Esperanto-bridged translation and so, I don't know which are the native languages of users. If if was enough to support the most spoken languages in the different continents, we could limit to this list : English, Spanish, French, Portugaise. To translate text from any of these languages to any other of the same list, you need 4 * 3 = 12 translation directions. Including Esperanto and using it for bridge translations, that would do 4 * 2 = 8 translation directions. The difference is not big. But if now we take 20 languages, we need 20 * 19 = 380 translation directions for direct translation, instead of 38 or 40 using bridge translation (depending on if the bridge language was or not in the previous list). I agree with what Jacob wrote before : bridge translation with Esperanto may be better than with English, but direct translation will generally be still better. For English, there are unoportunate homonyms like : right for the contraries of wrong and left, race for ethnic group or for running ... Esperanto may not have these kinds of homonyms, and the way words ends can determine if the word is a verb, a noun or an adjective, where in another language the structure a the sentence must be analysed to make a choice. Only for that, it may be one of the best possible bridge language. But even Esperanto does not support everything. For instance, only one way to put a verb on past, No 2nd person of singular, like "tu" in french or "du" in Swedish, only "vi" which is plural, like "you" in English or "vous" in French. According to ressources, I think 100 Mb for the whole pair is a bit under reality. My directory /usr/local/share/apertium directories is still 77 Mb with only 10 pairs. And according to sources directories after compilation, it is 534 Mb for the same pairs ! But I think the choice between using bridge languages or not can be done differently. If we were 1000 people or more to develop, up to the trunk level every combinations for 40 or 50 languages, that would be the best. If we were 100 people or more to develop, up to the trunk level translations for 40 or 50 languages from and to Esperanto, that would allready be very nice. In fact, if not far than 200 language pairs have been started in Apertium, around 60 % of them are on incubator and often left in that state. Really ready pairs are 5 times less. We do what we can. I don't know what I could have done before the next summer for eo->fr translation direction and when I will start working on wordlists in order to prepare a large fr-en bidix. I think it is the same for all of us even if student working on automatic translation may have more time to spend on Apertium. I don't know for what social network this request is about. But I thought to automatic translation for the website bewelcome.org putting in relation hosts and travellers. For that kind of website, between what would be nice to have now and what is really ready, there may be a factor of 10. So, bridge translations (and not only with Esperanto) may be temporary solutions waiting other pairs to be developped. And even like that, there will still be unsupported languages. But I think according what was done, Apertium is still the free translation software which has been developed the most. -------------------------------- Bernard Chardonneau (France) Phone : [33] 1 64 90 87 04 or [33] 9 72 36 32 90 (from Sept to June except holidays) GSM phone : [33] 6 49 95 13 95 (french scholl holidays, C zone) Multilingual websites for my free softwares : http://libremail.free.fr and http://libremail.tuxfamily.org http://cyloop.tuxfamily.org (mainly translated with Apertium) My general website (in french only) http://bech.free.fr ------------------------------------------------------------------------------ Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS, MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft MVPs and experts. SALE $99.99 this month only -- learn more at: http://p.sf.net/sfu/learnmore_122412 _______________________________________________ Apertium-stuff mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/apertium-stuff
