Re: [xml-dev] IDs considered harmful or why keys might be betterthanIDs...

["Jonathan Borden" <jborden@mediaone.net>]

Simon St.Laurent wrote:

>
> I think that perhaps you've misread my suggestion.

No, I was using it to make a point roughly along the lines of "perfect is
the enemy of good". In our rush to 'move beyond' DTDs we sometimes neglect
the lowly internal subset which _is_ instance data, so for example
specifications like SOAP which are now stuck with the inability to have IDs
(shrug).

> I'd like to throw
> away all of these approaches to external annotation of document instance
> content, not merely replace one with another.
>
> That'd leave us the document instance as primary, period.  No need to
> worry about internal vs. external subsets because some non-validating
> parsers (perfectly legally) don't bother loading external resources, no
> need to wonder whether the ID attributes in a document were identified
> using a DTD or an XML Schema.
>
> Instance supremacy, taken seriously, means never looking beyond the
> document. Limiting and liberating at the same time.
>

Right. And since we have relatively good information that every SAX parser
properly reports ID attributes _which are declared in the instance i.e. the
internal subset_ I am not entirely sure that the need for another syntax to
in-line declare identifiers is not syntax sugar (not that syntax sugar is
sometimes useful).

Jonathan