Re: Options for dealing with IDs

Elliotte Rusty Harold wrote:

> Steal undeclared attributes of name id (and ID). If an attribute is 
> called id, whether or not it is declared in the DTD or schema, and 
> whether or not it's type is DTD, it is treated as an ID for fragment 
> identifiers. However, no change is made to the attribute's type. All 
> parsers will continue to report the attribute's type as they currently 
> do, whether that's CDATA, ID, my-namespace:social-security-number, etc.

Hmm, the notion of "treating as an ID" is fuzzy enough that I don't want 
to go there.  All I'm suggesting is that for resource representations 
served as */xml or */*+xml, a fragment identifier of "xyz" points to the 
element which has the attribute id="xyz".

There's not really even any requirement that the value be unique, just 
pick the first one.  Worked fine for HTML for years. -Tim