El dt 09 de 10 de 2012 a les 21:01 +0200, en/na Per Tunedal va escriure: > Hi, > --snip-- > > > > Then do the magic and replace "Spanish" with "Swedish" and "English" > > with "Danish". It will work the same way. > > > > > (snip) > > > > > > > > > > > And do your algorithm on it. > > > > > > Weer should I build the algorithm? In a standalone module, or in some API? > > > What would be the hooks? I surely need to be able to get access to the > > > monodix'es and bidix > > > in some database form? > > > > No API, no databases. > > > > Do it in a standalone module. You have access to the output of the > > lexical-transfer stage as I described in my previous email. This is all > > the information you will need. If you want to get the information from > > the bidix and monodix in a text-format, you can use the "lt-expand" > > tool. > > Sure, the algo can initially be developed that way. It's better to make > small misstakes for a start. But when it starts to works smothly? Then > tests on a large corpus would be needed to fine tune it. How to do that?
I don't see any problem with running tests on a large corpus. Even with this ad-hoc method. We can revisit the problem when there are results. Also, I think it is better to discuss real problems rather than imagined ones. I quite like discussing imagined problems of space exploration, or the implications of a zero-scarcity society. But machine translation is something I actually work with right now. So I prefer to discuss the real problems that we have right now, rather than ones we might have at some undetermined point in the future. > > ...snip... > > No. > > Well, I would be happy to have those words added to the Swedish > dictionary in the pair sv-da. And it would be nice to keep the > possibility open to test the sense-selection-module if it seems > successful. How to proceed? You can proceed by adding the words. I'm not sure what is not clear about that! > Right now, I am preparing for using a new lexical selection module by > removing old LR/RL-tags (without adding slr or i-tags) and adding new > pairs when I run into them. I will be happy to try both Francis' and > Keld's module. Would be very interesting to compare different > approaches. Great! Fran ------------------------------------------------------------------------------ Don't let slow site performance ruin your business. Deploy New Relic APM Deploy New Relic app performance management and know exactly what is happening inside your Ruby, Python, PHP, Java, and .NET app Try New Relic at no cost today and get our sweet Data Nerd shirt too! http://p.sf.net/sfu/newrelic-dev2dev _______________________________________________ Apertium-stuff mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/apertium-stuff
