Well, what d'ya know? Thanks, Tim! I guess I'd better get around to
reading that spec.
At first glance, it looks like there are two ways of relating one
document/node to another. First, there is the LINK property in the header
of the node. There may be any number of LINK elements within the header,
each one specifies some relationship between the document and some other
specifed document. Secondly, a relationship can be specified for any
anchor using the REL or REV attributes. There are currently about 13
relationship types currently supported. See sections 3.1, 3.6, 3.15, and
chapter 6 of ftp://info.cern.ch/pub/www/doc/html-spec.ps.
So, why are people *not* using typed links? Has noone seen the need
or is everyone else as ignorant as I am? Is anyone out there creating
such links? Has anyone built support into their client(s)?
Why are typed links interesting? Well, one way they could be
immediately useful to me is in printing material from the web. For
instance, I have been reading about the web, HTML, SGML, HTTP, hypertext,
etc. The other night I wanted to take a hardcopy of some of the
documentation home with me, but I had to print it node by node. (I did
find the postscript documents available via ftp, but they just aren't
as current or complete as the material on the web.) If the links were
typed, and my client was smart enough, I could ask for a printout of
the current node and any anchors within it which are no more than 2 hops
away and whose type is "embed", "present", or "includes". Depending on
the particular node that I was at, this should get me a manageable amount
of information relevant to my current interest.
Of course, there are some design issues involved in deciding how to
display/present the results of such directives.
Chris
-----------------------------------------------------------------------
Christopher McRae mail: mcrae@ckm.ucsf.edu
UCSF Center for Knowledge Management at&t: 415/476-3577
530 Parnassus Avenue, Box 0840 fax: 415/476-4653
San Francisco, California 94143