1. The Web is more than big enough for many different formats to co-
exist, including SGML, HTML, PDF, probably also (shudder) RTF, etc.
2. In order to graft themselves onto the HTML web and take advantage of
its synergy, and get a significant synergy for themselves, organizations
promoting these other formats will still need to support URL/URI/URCs to
link to documents in dissimilar formats.
3. Therefore, any expansion of the semantics of linking in HTML should be
theoretically possible to support directly in other document formats.
Some, like HyperCard and I think PDF, already define things like 'Home'
'Back', 'Next', etc., so this level of hardwired semantics is all right.
4. What makes the Web the Web is links. Ultimately the formats that best
support navigation will win out. HTML may well be only a transition to
things more presentation-oriented (PDF or RTF), object-oriented (CORBA,
JAVA), semantically-rich and robust (SGML), or quite probably all of
the above. There is more than enough economic advantage to using any
one of these formats in specific applications, to support an industry
around each. The challenge here is to ensure that any link from one
format to another can retain the same semantics even if it is being
expressed in different languages. Another reason for extreme minimalism.
Otherwise we run the risk of a web that cannot be indexed in a format-
independent way. Argh! We'll be 'back' to keywords (where we are now)
not knowing which words out of the thousands on a page are significant.
I'd like to find out how Adobe is supporting links into/out-of PDF, look at
some HotJava stuff, etc., before deciding on a list of hardwired links for
HTML.
-- Craig Hubley Business that runs on knowledge Craig Hubley & Associates needs software that runs on the net mailto:craig@hubley.com 416-778-6136 416-778-1965 FAX Seventy Eaton Avenue, Toronto, Ontario, Canada M4J 2Z5