Re: AW: [all] revision to disambiguation text

Hi Yves,

On 18. 10. 2012 15:08, Yves Savourel wrote:
> Hi Tadej, all,
>
> Currently we have: in the global section of disambiguation:
>
> - Using disambigSource and disambigIdent to specify the collection and the identifier itself.
> - Using one of disambigIdentRef, disambigIdentPointer or disambigIdentRefPointer using a URI for the disambiguation target.
>
> It seems it should be:
>
> - Using disambigSource and one of disambigIdent or disambigIdentPointer to specify the collection and the identifier itself.
> - Using one of disambigIdentRef or disambigIdentRefPointer using a URI for the disambiguation target.
I declared disambigIdentPointer since I was going from the ITS1.0 
patterns, where this was intended to cover the case where one would 
declare the entity within the same document. I don't expect this style 
to be used often. Initially, didn't spot this ambiguity and the 
resulting side effects when we also allow a non-URI identifier. I would 
say that the second version makes more sense.

> Sorry, I'm being slow and having a bit of a hard time to understand what combinations of attribute are allowed. My understanding so far was that you could have:
>
> - class (and possibly granularity)
> - and either source+ident or identRef

Type class and granularity are optional. The mutual exclusivity only 
applies with the addressing modes (source + ident* vs. identRef*). Your 
example below is correct - the rules should say that one must use one of 
these two possibilities.

-- Tadej

>
> But the example 53 should a case with no class.
>
> The way the global attributes are defined currently is such that the disambiguation rule could have just the selector :)
>
> Are the possibilities:
>
> - Either source + ident (+ optionally claas) (+ optionally granularity)
> - Or just identRef (+ optionally claas) (+ optionally granularity)
>
> Or some other combinations?
>
> Thanks,
> -yves
>
>
>

Received on Thursday, 18 October 2012 13:33:02 UTC