[html4all.org] implementations and bug reports based on HTL 5 draft

[Robert Burns]

Hi Jason,

I'm not totally clear what your position is on these bug report  
comments. I agree with everything you say in this message, but when  
you get to the end you say:

> I am sure that all major implementors are aware of this.

I know Chris Wilson on the IE team is aware of this. I know that  
Chaals on the Opera team is aware of this. However, WebKit and  
Mozilla are much more decentralized communities. Patches and bug  
fixes come from all sorts of places.s. Someone reading through the  
comments on a bug may read these comments — deliberately written  
authoritative voice — and think they are authoritative. Yes it is  
true that upon further investigation they may notice the disclaimer  
in the document header. However, a programmer fixing a bug is usually  
not interested in such disclaimers. They diver in right to the  
section cited in the comment to find out what to implement.

So while I agree with what you said, this to me implies that members  
of our WG should never write comments that can only be intended to  
play with the ambiguity on this issue.  There's no reason for Lachlan  
to add a comment to a bug saying that HTML5 defines blank. when HTML5  
has not defined anything. the only relevant recommendation to cite at  
this point is HTML 4.01 or XHTML 1 or XHTML 1.1 on the issue. If this  
were a bug involving headers= and Lachlan wrote ""The headers  
attribute has been dropped from HTML5" in a bug comment, I think we'd  
agree that's not appropriate. Wouldn't we? To me, his comment  
regarding <input usemap>  is no different.

Take care,
Rob

On Aug 24, 2007, at 10:11 PM, Jason White wrote:

> The issue of bug reports being filed, and implementations possibly  
> adopting
> parts of the HTML 5 working draft at this early stage, has arisen  
> on several
> occasions during these discussions.
>
> W3C working drafts typically warn that the specification may be  
> changed at any
> time and that the working group will not be bound by early  
> implementations in
> deciding what changes to make. Early implementations are carried  
> out entirely
> at the risk of the implementor.
>
> Only late in the development process, and certainly by the Candidate
> Recommendation stage, are specifications considered stable enough  
> to be
> implemented.
>
> HTML 5 isn't even a W3C working draft yet, so it would be far too  
> early to
> make implementation decisions based on it, unless one is prepared  
> to undo or
> change those decisions as the draft develops.
>
> I am sure that all major implementors are aware of this.
>
> _______________________________________________
> list_html4all.org mailing list
> https://www.html4all.org/wiki