Hello Apertiumers!

I'd like to canvas opinions on creating a new top-level SVN module for
monolingual language packs. 

Apertium is primarily a machine translation project, but we also provide
substantial monolingual resources for many languages. For example, the
apertium-kaz and apertium-tat directories are currently in incubator/
but this does not fully reflect their status. They are quite
comprehensive morphological analysers and constraint grammars for the
languages in question. This also goes for apertium-gle (also in
incubator, for want of a better place).

The current top-level directory structure is as follows:

 branches/    branches of stuff in any of the other modules
 incubator/   stuff that has only just begun being developed
 nursery/     stuff that might be useful to someone
 staging/     stuff that is almost ready for release 
 trunk/       released pairs and software.

I think that insofar as monolingual language packs are concerned, they
don't necessarily belong in trunk/, it may be confusing for users, and
also, they could be standalone, not relying on all parts of the MT
platform. In any case, Apertium is primarily a machine translation
platform, and so the trunk (and the bulk) of the repository should be
dedicated to MT.

I propose adding a new top-level directory:

 languages/

This would have submodules like:

 languages/apertium-kaz (Kazakh morphological transducer + CG) 
 languages/apertium-tat (Tatar morphological transducer + CG) 
 languages/apertium-gle (Irish morphological transducer + CG) 
 languages/apertium-kir (Kyrgyz morphological transducer + CG) 

Language packs would graduate directly from incubator/ to languages/
upon reaching stable/mature status. Of course, packs would not have to
have a CG, they could equally well have a .prob file.

Having this separation would have further benefits: At the moment when
people want to use Apertium resources in their experiments/papers, they
have to cite a language pair. This would make it clearer. It would also
allow us to more easily and consistently package other people's
linguistic resources with an Apertium-flavour (As in the case of
apertium-gle).

I'd like to hear people's opinions on this, perhaps it is a question for
the PMC ? 

Regards,

Fran


------------------------------------------------------------------------------
Introducing Performance Central, a new site from SourceForge and 
AppDynamics. Performance Central is your source for news, insights, 
analysis and resources for efficient Application Performance Management. 
Visit us today!
http://pubads.g.doubleclick.net/gampad/clk?id=48897511&iu=/4140/ostg.clktrk
_______________________________________________
Apertium-stuff mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/apertium-stuff

Reply via email to