RE: Mechanism, not policy [was: Attribute uniqueness...]

>There is a fundamental philosophical shift going on, where the W3C
>published a recommendation, with a specific purpose, and now may be
deciding
>to retroactively change that purpose, despite the 18 months of use that
the
>Rec has gone through.

This is one reason I lean toward some version of Forbid/Deprecate. That's
the closest match to the original authors' intent.

The problem is that there appear to be at least two published specs which
_disagree_ on how to interpret the point under debate. At least one of them
is definitely going to take a hit. The question is only which, and how
badly.

A large part of this debate has been about which solutions cause the least
breakage. There doesn't seem to be an answer that causes _no_ breakage.

It would be nice if this hadn't arisen. It's a good reminder for us all to
be more careful about reviewing the interactions between specs, early and
often. But if we continue to run development cycles in web-years, this sort
of thing is likely to continue to arise.

"When you want something in the worst way.... you may get it in the worst
way."
______________________________________
Joe Kesselman  / IBM Research

Received on Thursday, 8 June 2000 13:25:55 UTC