Re: 2.3: Link Recognition by PI or other signal?

> From: Martin Bryan <mtbryan@sgml.u-net.com>
> >
> >2.3 Should we provide a PI or other signaling mechanism whereby a document
> >    can specify that particular elements ought to be processed as link
> >    elements?
> No - link identifying attributes should be all that is needed by any browser.

If we cannot have multiple attlists (in particular, an attlist in
the internal subset that declares only an element's "link identifying
attribute"), then it might be worthwhile to have some PI syntax that
allows the effective equivalent of an "attlist augmentation."  Then
this PI--which would look a lot like an attlist decl--could declare
the #FIXED link identifying attribute.

To restate: assuming we identify link elements via an attribute, I
still think we need some way to avoid requiring the link identifying
attribute to be set explicitly on every instance of this element type
in the absence of the complete DTD.  This capability seems best
provided by some sort of "partial declaration" such as an "attlist
augmentation" in the internal subset.  Now whether this takes the form
of a PI or not depends on what happens in the 8879 TC and what we do
with regards to multiple attlist decls in general in XML.