RE: I can't code, but...

> -----Original Message-----
> From: www-html-request@w3.org [mailto:www-html-request@w3.org]On Behalf
> Of Gordon Worley
> Sent: Thursday, January 07, 1999 10:08 PM
> To: Inanis Brooke; www-html
> Subject: Re: I can't code, but...
>
>
> Inanis Broke wrote:
>
> >A test suite sounds like a good idea, as long as it's "cheat
> proof." For me,
>
> 	This is part of why I would want the code for the browser, as I
> mentioned several e-mails ago earlier in this thread.  Cheating could
> possibly become an issue, since I would assume that the testing process
> would be avaliable for public review.  With this knowledge,
> Netscape could,
> for instance, make sure that the proper Java data bindings to make the
> OBJECT element work with specific mime types are built into the
> browser and
> work only during the W3C's test, keeping their certificate system in tact.

I think that is unrealistic. If something as you describe happened (and I
don't think the particular scenario you describe is remotely plausible,
given the logistics of typical OBJECT implementations), then
users/developers would quickly recognize that the browser only worked in the
narrow range of situations covered by the test, the proper folks at the W3C
would catch wind of it, it would be publicized. The inevitable bad press is
too much to risk.

> While I realize that it is unlikely that browser makers would want to
> implement features that would pass only the test, it would be necessary to
> be ready for all situations.

Having the code available wouldn't help anyone spot the problem--it would
only be available for someone to point out where the problem is once it's
already been discovered. And that's irrelevant to most people: the important
thing is that the problem is discovered, and the actual browser developers
(who will be fixing it) are the only people who really need to know the
precise nature of the problem at the code level.

Braden

Received on Friday, 8 January 1999 03:58:11 UTC