That short line got picked up, but nobody seemd to agree with me. But
suddenly everything has changed, REL is going to be standardized in an
HTML-independent way, and REV is going to be dropped. Nice work!
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)"
Furthermore, the value of REL will be something like:
scheme:/valuepart/valuepart/.../valuepart
(much like URLs, including the rules for relative values). For
example:
htmllink:/parent
htmllink:/next
htmllink:/x-uncle
previous *)
ToC *)
cooklink:/recipe/expensive
cooklink:/recipe/cheap/vegetarian
mathlink:/corrollary
mathlink:/example
x-proprietary:/opaque0/opaque1
*) For the moment, the implicit BASE is "htmllink:/", so that
"previous" expands to "htmllink:/previous".
|Dan writes:
|> This is very much the crux of the matter: are we defining a global
|> formal semantics, or just a global syntactic mechanism upon which
|> folks can build application-specific formal semantics?
Both: a global syntactic mechanism and a semantics for one corner of
that world. (The same is true of URLs).
Bert
-- Bert Bos Alfa-informatica <bert@let.rug.nl> Rijksuniversiteit Groningen <http://www.let.rug.nl/~bert/> Postbus 716, NL-9700 AS GRONINGEN