Hi Rodolfo, all,
>> It seems it would be simpler to have one element for each extracted
>> unit, and the un/segmented information inside that element: Any type
>> of parser can access all of it at once, no need for linking or other
>> extra mechanism; finding/fixing/editing files is easier; etc.
>
>
> Please keep unsegmented information separate from translatable segments.
> Like Andrzej, I don't plan to put unsegmented text in my XLIFF files.
> A common parent for unsegmented/segmented would be annoying unless it
> is optional.
>
> Unsegmented text should be optional, as it is not needed for translating
> a segmented XLIFF. Unsegmented text should not be in the core XLIFF
> module, it should live in a separate optional namespace.
Having the un-segmented text not part of the core and living in a separate optional namespace is a strange idea.
The first goal of XLIFF is to represent extracted text, not segmented extracted text.
The representation of the segmentation is what should be optional.
Cheers,
-ys