Ugh, I was as clear as mud.
* A. Pagaltzis <[EMAIL PROTECTED]> [2005-07-17 06:20]:
> It would make more sense to me to say that a particular type of
> @rel value always expects a dereferencable URI or never expects
> on. “in-reply-to” would be in the latter category.
>
> I can see the value in having both for cases where the
^^^^^^^^^^^ ie defining @idref and @href
> relationship may easily be either case. Although I don’t like
> the idea of both attributes being present at once; a Link
> Construct should refer to one resource, IMHO, and building some
> kind of fallback mechanism into it rubs me the wrong way for
> reasons I can’t quite verbalize yet.
>
> In which case you could add another attribute to cover this.
Ie rather than @idref, a boolean attribute that says whether
@href is dereferecable or not.
> Or you could decompose it into two relationship types, both of
^^ referring to this boolean attribute
> which mean the exact same, but of which one expects a
> dereferncable URI and the other expects an abstract one.
Regards,
--
Aristotle Pagaltzis // <http://plasmasturm.org/>
