LINK and META extensions as prototyping technology


Daniel suggested you were unsure what I felt was experimental
about the LINK and META suggestions.

I view the revised LINK and META capability as permanent features
of the language that should be hardwired into the DTD starting
with 4.0.

I view the rather general metadata capability that the expanded
LINK and META elements would give us as a fitting technology for
markup experimentation.

	+ It is very capable.  You can say almost anything.

	- It lends itself to obscure manners of expression,
	less direct than what you look for in HTML attributes.

People will be less motivated to use it in general than the same
semantics standardized in HTML attributes, because the semantics
of the user-defined keywords was not guaranteed by the standard.
But people who want to experiment with specific functionality
can develop prototype sites and client functions that used the
information model of their choice by agreement between the server
and client experimenters on keywords.  They don't have to violate
the HTML DTD to do this, if the user-defined LINK and META are
capable enough.

So I would see permanent rules for LINK and META that stay the
same in HTML, and transient/experimental uses of LINK and META
for experiments and gap-filling with regard to the semantics of
the standard markup defined in the HTML specification.

Functions that prove to be winners would likely be the subject of
proposals to add features in HTML.  Functions that fail would not
put HTML through the throes of a deprecation cycle.

-- Al

PS: META is probably gravy in this regard because we have CLASS
very widely available.  The trick is being able to put REL, REV,
CLASS and the like under strict definitions imported from an
external lexicon.  That requires being able to target a
dictionary LINK to an individual LINK element, for full

Received on Tuesday, 14 October 1997 15:38:40 UTC