Re: W3C press release: comments on standard-setting

[David Brownell <david-b@pacbell.net>]

> Recommendation has always been the end of the line (those who think we 
> should give up the ghost and just use "standard" can now fight with those 
> who are appalled at the W3C as a body deeming anything standard).

I'd think "end of the line" would be "obsolete, don't use in new software".
So I'll disagree about that point ... :)

Anyway, it's clear there are errata-tracking and other spec upkeep processes
that happen after REC, even if W3C doesn't emphasize them ... I think a "real
standard" (vs W3C REC) will normally have had those later processes applied.


> The Candidate Recommendation was supposed to be "we think it's cooked, go 
> out and tell us if we're wrong", but the vague manner in which CR was and 
> still is defined has turned it into a "last, last call" instead of a 
> serious or lengthy implementation and kick-the-tires phase.

Some of us think that the W3C processes change too often to rely on
for any real standards work.  That "implement-and-kick-tires phase" is
a pretty new concept for W3C, and as you pointed out it's not being
applied universally.

I think a more consistent "implement and abuse" experimental phase would
be a good general policy.  In fact, I think many folk here would likely agree
that for some specs, that process doesn't really start until REC ... which is
another reason the "post-REC" part of the W3C spec lifecycle may be worth
some additional (scientific :) scrutiny.

Some folk have suggested that the natural completion of the W3C process
would be like a butterfly breaking its chrysalis ... to fly as a real standard,
with full emphasis on conformance and interop between implementations.
(And maybe hosted by some accredited standards body!)

- Dave