Re: DOCBOOK: Linking in DocBook (specifically for EBNF, but more generally as well)

["Eve L. Maler" <Eve.Maler@east.sun.com>]

At 09:43 AM 4/10/00 -0700, Terry Allen wrote:
>So use of an xlink: attribute specifies the semantics of the link
>but not the syntax?

I'm not sure what you mean.  It certainly specifies XLink-standardized 
semantics, and it also imposes some constraints on the syntax.  E.g., 
currently the only way to supply a URI to be used in an XLink-based way is 
with xlink:href (where "xlink" can be any prefix, as long as it's bound to 
the right namespace URI).

>Which is not to say you can't use another method, if you define it
>and don't use URLs.  (Note that the URN WG did not agree that fragment
>identifiers made sense with URNs, so there is no provision for them.)
>Not that I'm suggesting it; the point is that the semantics of the
>fragment ID are bound to the media type of the target, so you can't
>make up your own syntax for fragment IDs (although queries are fair
>game ...).

Yep.

>"#foo" alone would be relative to the current document (whatever
>that will end up meaning).

Yes, but "" along would also be relative to the current document. :-)  It's 
independent of the presence of the fragment ID.

>I think I've got it now.  But I'd like to switch over consistently
>throughout the DTD, and when tool support is ready.

Yep, I can see that.

         Eve

Eve Maler                                    +1 781 442 3190
Sun Microsystems XML Technology Center    elm @ east.sun.com