> It certainly would make some tasks of
> working with XLIFF much easier if the
> skeleton were stored at the unit level.
> From a conceptual standpoint, I think it
> is much cleaner than other options. I think the cleaner option is to have no skeleton in the XLIFF document :) Just a pointer to it. Or even nothing at all (and let the tool work out the link using its package/manifest). The question is: do we want XLIFF 2.0 to have an optional *standard* skeleton mechanism? (There is no such requirement in the feature list currently)
If the answer is yes, then having a place to store the skeleton between units would certainly be a clean and simple way to do it. Note that it won't work easily with all file formats (try this with an IDML file for example, or a .NET satellite assembly). Cheers, -ys
