Hi, a merge would make the pair is-sv benefit from the continuing corrections, improvements and additions to the apertium-swe.swe.dix.
I've taken a look at those dictionaries and found that the differences are considerable. The first impression is that Tihomir has made the dictionary apertium-is-sv.is-sv.dix very much more neat and tidy. I like his grouping of similar items. That makes it more easy to maintain the dictionary and to assure that everything in a category is included. (I have started to do something similar with pronouns, because that have been a mess. I'm still not sure if all pronouns are included in the dictionary. It needs to be checked.) The words in apertium-swe.swe.dix are still mainly grouped after how/when they have been included. This is due to an ambition to indicate the relative need for checking. Maybe this can be changed when more words have been checked and corrected in some distant future. Second: The months are treated differently. I'm not sure exactly what Tihomir has done, but it looks neat. This applies to the bidix as well. Can anyone explain? Should this be used in apertium-swe.swe.dix and apertium-swe-dan.swe-dan.dix ? Third: A merge presupposes changed terminology for pronouns in apertium-is-sv.is-sv.dix: subj = subject form changed to the traditional nom = nominative obj = object form changed to the traditional acc = accusative That would be a major change. I suppose this would deteriorate the translation quality due to an immediate need to retrain the tagger, just as for the pair sv-da. All the same such a change would make the is-sv pair compatible with other pairs, and open up for new language pair like is-da and is-no. Forth: If any words in apertium-is-sv.is-sv.dix still aren't present in apertium-swe.swe.dix, it would be easy to include them. Yours, Per Tunedal ------------------------------------------------------------------------------ Dive into the World of Parallel Programming. The Go Parallel Website, sponsored by Intel and developed in partnership with Slashdot Media, is your hub for all things parallel software development, from weekly thought leadership blogs to news, videos, case studies, tutorials and more. Take a look and join the conversation now. http://goparallel.sourceforge.net/ _______________________________________________ Apertium-stuff mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/apertium-stuff
