>[ burchard@horizon.math.utah.edu proposed HTML needs an abstract link model]
I agree we need an abstract link model, something *simple* and extensible.
The simplest link type is just a string that is associated with the link,
something like the C typedef, it doesn't do anything but give it a name.
Various hypermedia systems including NoteCards, Neptune, KMS, and Intermedia
have abstract link models... the least common denominator of them all are
user-named types with no software-interpreted semantics... next most common:
free form attributes, then various means of automatic traversal/map-making
(e.g. mandatory 'contains' links, 'citation' and 'author', etc.), which were
more or less practical depending on the coherence of the authoring community.
Ian Graham's list of links is a reasonable starting point as it grows from
practice in other Internet-based communities. Murray's description of SCO's
work in online docs sounds intriguing, and may provide a few more key pieces.
Not to mention SCO has the only real experience with HTML REL/REV as it exists.
If SoftQuad will move forward and be second, that would help immensely.
As all the people named here save one are within a few blocks of me here in
Toronto, and since Murray and I have the most diametrically opposed positions,
(which appear to me to be reconciled by taking the list of types out of HTML),
and since he represents a vendor company, and I a set of authoring communities,
and since in a previous life he was working with CALS standards, and I with the
local-area hypermedia systems (and later the first web browsers back in '91 or
whenever that was...!), I think maybe he and I should author the draft together.
Whatever we can agree on, will probably be both minimal and necessary... as a
standard should be...!
--
Craig Hubley Business that runs on knowledge
Craig Hubley & Associates needs software that runs on the net
craig@passport.ca 416-778-6136 416-778-1965 FAX
Seventy Eaton Avenue, Toronto, Ontario, Canada M4J 2Z5