Hi all, I think the task
"find X rules for how to translate words with more than one possible translation" could be misunderstood as they could mixed lexical selection problems with part-of-speech ambiguity problems. I see graduate students doing so, every year. Cheers -- Felipe El 16/11/11 15:20, Francis Tyers escribió: > El dc 16 de 11 de 2011 a les 14:18 +0000, en/na Jimmy O'Regan va > escriure: >> On 16 Nov 2011 14:09, "Francis Tyers"<fty...@prompsit.com> wrote: >>> >>> El dc 16 de 11 de 2011 a les 14:02 +0000, en/na Jimmy O'Regan va >>> escriure: >>>> On 16 Nov 2011 13:35, "Francis Tyers"<fty...@prompsit.com> wrote: >>>>> >>>>> Hey all! >>>>> >>>>> I've thrown all the parts together and have a working prototype >> of >>>> the >>>>> lexical selection module. A rule compiler, and a processor. >>>>> >>>>> At the moment the rule format is like: >>>>> >>>>> >>>> >> https://apertium.svn.sourceforge.net/svnroot/apertium/branches/apertium-lex-tools/examples/rules.txt >>>>> >>>>> But we have also discussed an XML-based format, which would be >> like: >>>>> >>>>> >>>> >> https://apertium.svn.sourceforge.net/svnroot/apertium/branches/apertium-lex-tools/examples/rules.xml >>>>> >>>>> I would like to, as my next step, improve the rule compiler (at >> the >>>>> moment there is a lot of string mangling that I think could be >>>> improved >>>>> on -- e.g. for holding the pattern lengths/ids), and support the >> XML >>>>> format, but in order to do this, I would first like to get >> comments >>>> on >>>>> it. Is there anything that you would change? Do you feel >> comfortable >>>>> writing rules in this format? >>>>> >>>> >>>> It might be better to ask next week, when GCI tasks have been >> sorted >>>> and finalised. Split focus and so on. >>> >>> What a great idea! We could make some GCI tasks like "come up with X >>> lexical selection rules for a language pair of your choice". >>> >> >> You'll want to rephrase that, significantly. GCI students are casually >> browsing a list of titles so you should pick a title that doesn't rely >> on a relatively obscure phrase - something that immediately informs >> them that they probably already know this. > > Yeah, how about: "find X rules for how to translate words with more than > one possible translation" ? > >>> I was also thinking of a GCI task to make a human interface to >> creating >>> the rules (maybe web-based?), offering the user all the possible >>> sentences, and asking them to mark them as "ok/not ok" and mark the >>> "contextually important words". I can't think that it would be more >> than >>> a couple of days work for someone with experience with PHP. >> >> Nice idea. There are plenty of javascript draggy droppy things around, >> so we could maybe also look for an interface to build basic transfer >> rules too. > > Exactly, also, web 2.0 stuff is hip with the kids! > > Fran > > > ------------------------------------------------------------------------------ > RSA(R) Conference 2012 > Save $700 by Nov 18 > Register now > http://p.sf.net/sfu/rsa-sfdev2dev1 > _______________________________________________ > Apertium-stuff mailing list > Apertium-stuff@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/apertium-stuff -- Felipe Sánchez Martínez Dep. de Llenguatges i Sistemes Informàtics Universitat d'Alacant, E-03071 Alacant (Spain) Tel.: +34 965 903 400, ext: 2966 Fax: +34 965 909 326 http://www.dlsi.ua.es/~fsanchez ------------------------------------------------------------------------------ All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d _______________________________________________ Apertium-stuff mailing list Apertium-stuff@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/apertium-stuff
