Re: REL and REV attributes (Was: More comments on HTML 3.0)

Daniel Shields (pepprboy@inforamp.net)
Tue, 2 May 95 02:57:45 EDT

>Bert Bos <bert@let.rug.nl> wrote:
> If I interpret the last messages by Dan Connolly and Craig Hubley
> correctly, then it seems that REL is the functor of a 2-place
> predicate, the 1st arg of which is the current document, the 2nd is
> the document in the HREF-attribute, or symbolically: "REL(BASE,HREF)"

If I understand everything leading up to this correctly, then REL/REV
attributes specify knowledge-base facts *about* the links rather than
functions to be performed on them. Where, for example, if HREF="place"
REL="fact" then fact(THISDOC,place), conversely if HREF="place" REV="fact"
then fact(place,THISDOC). The potential of this knowledge-base "web" should
be explored before it is discarded because, I feel, it has been dangerously
ignored.

So far, much of the discussion about this issue has been about what is the
meaning of links. I feel that we are on the edge of realising that the
current vision of hypertext on WWW is insufficient, and it appears that we
are now invoking 'kludges' in order to enrich our notion of hypertext.

We should be very careful to delineate specifying markups within documents,
from specifying the generalised feature-set of hypertext systems. IMHO we
should be prepared to handle, in the not-too-distant future, documents which
make references to multiple hypertext systems within the same document. But
are we ready to include references to arbitrary or unknown hypertext systems
within a generalized link syntax?

There is still something to be said about storing information about
documents in a parallel hypertext system. Information contained in a WWKB
(a world-wide knowledge base), of information about documents, is more
accessible as a distributed database, than it is when it is encoded within
the headers/bodies of the documents themselves. From a text processing
standpoint, this model is quite a bit more efficient.

As above, in delineating the hypertext system from the markup system, we
should establish only what is relavant link-information about a document,
and store that within the document. Information about the handling of sets
of documents by server systems, belongs in a parallel hypertext system
altogether. For instance, a document authority should be told about the
hierarchy of a set of related documents, rather than that information being
stored (redundantly) within the documents themselves.

--
Daniel Shields                                          #1054042 Ontario Inc.
(416)597-8239                                operating as Daniel Shields Ltd.