Re: Input to Best Practice 18: Assign unique identifiers to text items when possible

Hi Yves,

Yves Savourel wrote:
> Hi all,
>
> I'm working on my AI (http://www.w3.org/International/its/track/actions/21)
>
> One of the change to do is:
>
>   
>> We should align this formulation
>> "Your DTD or schema should provide xml:id (or an equivalent mechanism)"
>> with BP 9, e.g. like:
>> "Your DTD or schema should provide xml:id (or a different attribute to 
>> be of type ID)"
>>     
>
> I would tend to disagree: why be more specific in this Author BP than the others? In BP 15, 16 and 17 are not specific about what
> exactly means 'equivalent'.
>   

Probably the case of  BP 18 is different than BP 15, 16 or 17: in the 
latter three BP, we assume different markup with the same 
functionalities. In BP 18, however, the functionalities of unique 
identifiers can be rather different (although all share the purpose of 
unique identification). See as an example of a different functionality 
the mention of xs:unique/xs:key by Jirka at 
http://lists.w3.org/Archives/Public/public-i18n-its/2007JulSep/0075.html

> The authors don't really give a vole's pattouti about what are the specifics of the equivalent mechanism (or of xml:id for that
> matter).
>   
Probably the differences between its:translate vs. e.g. the "dita" 
translate attribute (see BP 17) are not so large, compared to 
differences between xml:id vs. xs:unique/xs:key . Hence, we thought that 
in BP 9 and BP 18 the alignment and actual repetition of material is 
rather helpful to the reader.

Felix

Received on Wednesday, 19 September 2007 06:06:41 UTC