W3C home > Mailing lists > Public > www-html@w3.org > November 2003

Re: XHTML 2.0 User Agent Conformance

From: Jim Ley <jim@jibbering.com>
Date: Sat, 1 Nov 2003 12:33:54 -0000
To: www-html@w3.org
Message-ID: <bo0999$oti$1@sea.gmane.org>


"David Woolley" <david@djwhome.demon.co.uk> wrote in message
news:200311011020.hA1AKHj06410@djwhome.demon.co.uk...
>
> > > Have you considered the cost of shipping a conformant browser?
>
> Shipping a corrected browser will cost something similar to shipping
> one with a security hot fix.  Shipping such versions is common these
> days!!

I don't agree with this, the regression testing on a security bugfix is only
relevant to other security holes, if you have to do regression testing to
ensure total 100% compliance to the specification (remember I wrote in
response to a suggestion that went beyond the WF requirement to actual
document validity) then the cost of the fix is higher.

> All new W3C "standards" have to demonstrate two interoperable
implementations
> (at least of any given feature).

Unfortunately though this requirement is not really sufficient, since it can
mean two features which are impossible reconcile in a single viewer still
make it into the specification so long as there are 4 viewers implementing
part.

I personally would like to see a requirement that there also be two
implementations covering a major portion of the specification, but I fear
this would probably lead to stagnation in the specifications,  I'm not sure
of the solution, but the interopable implementations is an excellent start -
we need to ensure it is stood by.

Jim.
Received on Saturday, 1 November 2003 07:43:08 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 23:40:09 UTC