Subject: Call for dissent: csprd02 148, allow empty group elements (RE: [xliff] Should group elements be allowed to be empty?)
Regarding csprd02 148 I propose we allow empty group elements (change “One or more <unit> or <group> elements in any order followed by” to “Zero, one or more <unit> or <group> elements in any order followed by”).
If I do not hear dissent by the end of this week, I will assume this is approved.
Good point Yves. We would just have to assume it would be unlikely that somebody would want to create an XLIFF file for a source which there was no translatable text (oh, maybe a use case could be imagined where the Resource Data Module is being used).
Just like the strict XHTML DTD (http://www.w3.org/TR/xhtml1/dtds.html#a_dtd_XHTML-1.0-Strict) says the following is perfectly valid:
I cannot imagine why somebody would write such a thing, but nothing prevents it.
Just an extra note: if we go with allowing empty groups, then we could have a <file> without unit as well, since a <file> can have one <group> only.
I have no opinion.
In XLIFF 1.2 we allowed empty group elements. In XLIFF 2.0 we require group elements to have one or more unit or group elements.
When I read the first sentence in the definition of group, “Provides a way to organize units into a structured hierarchy,” I thought, okay, not allowing empty groups makes sense. But when I read the second sentence, “Note that this is especially useful for mirroring a source format's hierarchical structure,” I became less sure. I think you if a writer’s goal for group is to mirror a source format’s structure, and part of that structure is a non-inline empty element, it would be reasonable to have an empty group element. The rub is we have inline elements meant to mirror empty source elements, but we do not have structural elements to mirror empty source elements. An example the comes to mind is the CALS table colspec (http://www.docbook.org/tdg/en/html/colspec.html ).
I think it is implicit in the fact that we are making statements that the group element can preserve structure and hierarchy of source files – that we intend to continue to support the maximalist (storing structure w/o skeleton) method. And I think we should.
So I think we should either:
2. Edit the sentences about preserving structure and hierarchy away from all elements except for skeleton.
I vote for 1.
Please let me know if you disagree. If nobody objects on this thread after a while, I will elevate this to a call for dissent.