Re: Is "Web Applications 1.0" a sensible starting point for the next iteration of HTML ?

Tina Holmboe wrote:
> On Mon, Apr 23, 2007 at 11:39:19PM +1000, Lachlan Hunt wrote:
> 
>>> Let's not mix the cards? The WHATWG proposal may, at some point in 
>>> the future, be adopted by the W3C's HTMLWG.
>> Your whole argument is based on a minor technicality, and that makes 
>> your point rather moot.
> 
>   It does not, no. You may disagree, of course, but I don't consider
>   this a "minor technicality".

Please explain why the fact that it is not yet officially hosted at the 
W3C is not a minor technicallity, given that the spec is very highly 
likely to be adopted by the HTML WG within a week or two?

It is a minor technicallity because you arguments are not based on the 
actual merits of the spec itself, but rather the credentials of the 
organistaion that created it.

>>> Very well: the WHATWG WA1 specification is not today, and MAY not be 
>>> tomorrow, the base for HTML5,
>> At this stage, it would seem rather naive to expect otherwise.
> 
>   That is very possible. 
> 
>   However, and luckily, HTML 4.01 isn't going anywhere.

For implementers today, HTML 4.01 is virtually irrelevant for the real 
world.  The fact is if you were to implement HTML 4.01 as defined 
(ignoring the fact that it isn't actually well defined), you would not 
end up with a browser that is capable of rendering the vast majority of 
web sites on the web today.

>> Based on feedback from both implementers and other contributors, it is 
>> indeed very likely that some parts of the spec will be significantly 
>> altered between now and its completion.  Again, I fail to see the 
>> relevance of your point.
> 
>   Implementing, re-implementing, and then re-re-implementing is only
>   a good idea in ivory towers

Acutally, the approach I described is anything but an ivory tower 
approach.  The ivory tower approach would be closer to what you seem to 
be advocating where the spec is written and only once it's stable should 
implementers start implementing it.  That simply does not work in 
practice.  XHTML2 is a good example of a spec developed with the ivory 
tower approach, which will see very little implementation and use in the 
real world (at least not on the web, perhaps in intranet apps or a 
walled garden).

> The implementation you speak of should /start/ when a solid foundation exist.
> 
> The WA1 document is /not/ a solid foundation for implementation at this stage.

It may not be totally solid yet, but it requires implementation feedback 
to be improved and become implementable.  Implementers should indeed 
start implementing features as soon as possible so as to provide such 
feedback to improve the specification.

>> seen very little use in reality have not.  I do not understand why it 
>> would be sensible to begin with a spec of such poor quality as the basis 
>> for work on the next version of HTML, particularly when a significantly 
>> better spec already exists.
> 
> A significantly better specification does /not/ exist.  Not yet. The
> current HTML 4.01 specification would, IMHO, be as good, if not 
> better, starting point than the WA1 draft.

Please provide some evidence to support your claim that HTML4 is a 
better spec than the HTML5 draft.  You could, perhaps, pick a few 
sections in HTML4 and explain why they are better than the equivalent 
sections in HTML5.

> And we already /have/ implementations of it.

Please provide evidence of a conforming implementation of HTML4, 
preferably a web browser that is in use in the real world.

> Actually, reality has not proven that at all. What /is/ proven is that
> UAs add things they'd like to see, or that authors clamour to use,  As
> a final, afterthought, a specification is written.

That's called market forces and it helps in determining what is and is 
not worth defining.  Specs that ignore reality are generally doomed to 
failure.

>   This is how we got IMG,

Are you suggesting that img is conceptually bad, or just that it has 
known design flaws?

> doesn't change the fact that *canvas might not be a very good 
> idea*.

Do you have any specific objections to canvas, or is that just a 
hypothetical example?

>> As it is a working draft, of course it isn't perfect.  But do you have
>> any significant issues with it that you would like to see addressed? 
>> Perhaps you should raise those issues on public-html or the whatwg 
>> mailing list.
> 
>   Once it becomes the WD, then I shall certainly give input on
>   the public-html mailing list. I am, however, not going to join
>   the *WHATWG* mailing list.

So it does seem that you are indeed judging the spec based upon where it 
resides, rather than it's actual merits.

>> You have not presented any arguments against the spec, except for the
>> fact that it isn't hosted at the W3C.  But even that argument is
> 
> I have presented several.

Really?  All your arguments seem to stem from the fact that it's not yet 
hosted at the W3C.  If you could point out an argument I may have missed 
which does not, I would appreciate it.

> In my not too humble opinion the WA1 draft is not of sufficient quality
> to be chosen as the starting point of a new revision of HTML

You have not yet indicated a significant deficiency in the HTML5 spec to 
support your claim of poor quality.

-- 
Lachlan Hunt
http://lachy.id.au/

Received on Monday, 23 April 2007 15:58:17 UTC