Subject: RE: [xliff] Extensibility in Modules
Thanks David for the reply. Of course, you are right on all points. Playing devil’s advocate, I just think the spec itself can be interpreted differently since it does not clearly state what you just stated below. *Simple* and *Glossary* can be interpreted quite different from *complex* and *terminology* in my dataset. In any case, we will support <mda:metadata> on <gls:glossary> as an extension in true spirit of standards and interoperability. And I’m not worried about it being removed, although some implementers might. Again, it’s too bad that there isn’t consistency in the spec on how to deal with <mda:metadata> as a module or as an extension overall, not just in <gls:glossary>.
From: Dr. David Filip [mailto:David.Filip@ul.ie]
Ryan, I don't want to be dogmatic here
But if you want to transfer terminology in your package, you should use the gls as the baseline to promote interoperability. Other people will be able to use at least the simple goodness without building support specifically for your terminology in mda.
If anytime you think are more feature rich than what is in the module, you are going to replace it with a proprietary mda construct, we are going again down the XLIFF 1.2 extensibility route
mda although a module is just an extension mechanism that does not allow for specification of any business logic
No one can really consume your mda based terminology as terminology
gls is tied to the core annotation mechanism and can be used for term lifecycle mgmt while terminology is in the XLIFF facilitated roundtrip
Outside of the XLIFF roundtrip terminology is supposed to live in TBs i.e. outside of XLIFF scope..
I think that the gls has a basic terminology roundtrip logic (tied to core) that gives you a terminology specifc baseline that the mda can never express.
All your structures that don't fit the gls you can transfer from your core extension point to the gls extension points.
If you feel that the data isn't safe there, you can suggest explicitly listing mda at those gls extension points.. I think this would be a minor change that would cause no harm.. would work for any XLIFF 2, and from 2.1 it would be treated as a guaranteed module.
On Tue, May 19, 2015 at 9:24 PM, Ryan King <email@example.com> wrote: