Re: LINK Element Confusion

I write, about using the PROFILE attribute of HEAD to specify a file
defining supplemental linktypes:

> >However, as noted by the Web Design Group
> ><http://www.htmlhelp.com/reference/html40/head/head.html>, there is no
> >format specified for this profile.  Does anyone know of a format
> >that's in use?

Frank Boumphrey replies:

> I don't know of a format. It may be a good idea to work on one, because as
> the  concept of the web as a series of objects grows, I think every one in
> this thread agrees  the LINK element provides a suitable way of describing
> relationships.

> I suppose a file with a SGML type comment would be as good a way as any

> <!--supplemental values for % LinkTypes
> 
> PARENT
> CHILD
> CHILDLIST
> NOTES
> -->

	I'd add HOME and MADE to that list.  (What's the meaning of
CHILDLIST?  A list of the document's children?  Seems a little
redundant.)

	The format should also include brief descriptions of the
meanings of the new link types, akin to what's in the W3C's
recommendation.

> Getting back to the method of defining values for I suppose the best way
> would be to use SGML syntax

	It seems like something more sophisticated than comments could
be a good idea, but I don't know enough about SGML to propose it.
Something like

<!ENTITY % LinkTypes "
   ALTERNATE     -- alternate version --
   | STYLESHEET  -- external stylesheet --
   | START       -- first document in the current collection --
   | NEXT        -- next document in the current ordered sequence --
   | PREV        -- previous document in the current ordered sequence --
           -- etc --
   | NOTES       -- notes or footnotes --
   | PARENT      -- logical parent --
   | CHILD       -- logical child --
   | HOME        -- logical ultimate anscestor --
   | AUTHOR      -- author or target for feedback --
   | MADE        -- converse of AUTHOR --">

(I make no promises about the correctness of my SGML syntax!)  I think
it's okay that this would seem to contradict or override the
definition of LinkTypes in the DTD, since this profile is not supposed
to be merged with the DTD, but rather define the conventions used in
the HEAD.  So any CDATA is legal as a linktype from the point of view
of the document itself, but authors should restrict the actual values
to what's defined in their PROFILE to aid automatic indexing.  Once
some conventions were set down, standard choices of PROFILE could be
stored somewhere central, and bots could recognize the protocol used
from the URI and not have to retrieve the list of link types.

	Are the MetaData people doing something like this as well?

					John T. Whelan
					whelan@iname.com
					http://www.slack.net/~whelan/

Received on Thursday, 20 August 1998 14:45:43 UTC