Re: AW: [all] revision to disambiguation text

Hi Yves,

On 23. 10. 2012 19:56, Yves Savourel wrote:
> Tadej,
>
> I thought I got it, but then the examples don't match my understanding (or vice-versa):
>
>> When specifying the type class:
>> * disambigClassXYZ mandatory
>> * disambigGranularity optional
>>
>> When specifying the identity:
>> * exactly one of:
>>    ** disambigIdentXYZ + disambigSource
>>    ** disambigIdentRefXYZ
>> * disambigGranularity optional
> The spec also says: "When using a disambiguation rule, the user MUST use one of the use cases for disambiguation: specifying the target type class, or specifying the target identity."
>
> But the example 52 seems to do both:
> file:///C:/MLWebLT/root/WWW/International/multilingualweb/lt/drafts/its20/examples/xml/EX-disambiguation-global-1.xml
>
> There we have its:disambigClassRef and its:disambigIdentRef used at the same time.
> So both the target type class and the target identity are define.
> But the wording "... user MUST use one of the use cases" seems to indicate either one must be used, not both.
>
> Can they both be use at the same time?
>
> If so we should have something like this:
>
> * At least one of the following:
>     * To specify the type class:
>        * disambigClassXYZ mandatory
>        * disambigGranularity optional
>     * To specify the identity:
>        * exactly one of:
>           * disambigIdentXYZ + disambigSource
>           * disambigIdentRefXYZ
>        * disambigGranularity optional

I thought about this for a while, and think that the mutual exclusivity 
implied in the text may be too strict. I would be in favor of allowing 
both to be specified, since it's pretty common to have both: in some 
implementations that I've seen, the type classes are detected in one 
pass, and the actual identification is done later on separately.

-- Tadej


>
> -yves
>
>
>

Received on Thursday, 25 October 2012 14:25:40 UTC