Re: brainstorming: test cases, issues, goals, etc.

Mihai Sucan schreef:
> Personally, I believe that the ones in charge of the HTML WG should 
> take the lead and come out with proposals of "what's next". This could 
> be done like a survey:
>
> 1. Start from scratch or continue work from WHATWG HTML5?
>
> If work is to be continued from WHATWG HTML5: due to the patent 
> policy, everything must be reviewed. Can this be done - is it a 
> feasible plan? If not, what else can be done? Is it possible to have a 
> "blanket patent policy" applied over all previous work? If not, it's 
> going to take lots of time.
>
> 2. Should the Forms Task Force start from scratch or continue from Web 
> Forms 2?
>
> This is very similar to the above issue, however we already have an 
> implementation in Opera, and some people started using it (I do). The 
> FTF should continue the work and make WF3 - the name is irrelevant. 
> XForms needs to be taken into consideration - "best of both worlds".
>
> Is the patent policy issue solved for WF2? IIRC, contributions were 
> made via the WHATWG mailing list without prior signing of any patent 
> policy.
>
> I don't think starting from scratch is at all good, in both cases 
> (HTML5 and WF2).
>
> As I understand, the working group needs to reach common grounds 
> before any of these important decisions are made. Could the "vote" of 
> each WG member be tracked? Perhaps a new web survey?

It’s important to get the correct questions first, before asking for 
answers :).

I’d say the situation is as follows; the three ‘alternative’ browsers 
represented in the WHATWG are very eager to implement new stuff. There 
are some nice specifications out there. However, for a new formats to 
really succeed, you need it implemented in Microsoft’s Internet Explorer 
as well, otherwise it will always only be used in a niche market (just 
like, say, XHTML is now), unless the browser market changes really 
significantly over the next few years. And the patent policy is an 
important prerequisite for Microsoft, especially after the Eolas patent 
case. Therefore, I think it’s inevitable that the W3C should take the 
lead here, and not the WHATWG, if the specification is meant to succeed.

So indeed, reviewing is a lot of work. Starting from scratch isn’t 
exactly appealing either. So, instead of posing people with two 
black/white choices, the interesting question is, what is there 
inbetween those that may be an interesting approach? After a little 
thinking, I came up with two:

1. A continue-the-work approach within the limits of the patent policy.

For something to be patented, it must be non-trivial, and there must not 
be prior art, right? One could run through the WHATWG spec and look for 
those things first. This would include things various structural markup 
elements such as the <section> element, <figure> etc, which is pretty 
trivial, and the Web Forms 2.0 forms validation which has plenty of 
prior art. If a person who has some reasonable knowledge of what 
software patents are out there in the www market would do this, I think 
you could very quickly get a significant part of the specification out 
of the way. The pieces that remain (say, the <canvas> element) can then 
be approached on an individual basis, where either the HTML WG members 
suggest a better approach, or the original ‘inventor(s)’ is tracked down 
and asked to sign a release form. I think this would allow a fair part 
of the WHATWG spec to be copied verbatim, and create some healthy 
discussion points for the remainder.

2. A start-from-scratch but at the same time reuse-WHATWG-work approach.

The W3C could make a list of key features of the WHATWG spec, and put 
them up for discussion individually. Participants also propose features 
they would like to see. For a lot of items, consensus will quickly be 
reached and implementation details can be copied from the WHATWG spec 
without much patent danger. For other items, again, healthy discussion 
takes place. This will make it easier to review the WHATWG spec’s 
features individually on their own merit, just as if they were 
well-worked-out proposals of HTML WG participants, and at the same time 
new ideas can be proposed on an equal field, and independently of an 
existing specification. As opposed to having a fairly finished spec out 
there where you have to fight a virtual boxing match when you propose 
any significant changes.

Any other suggestions?

Now, with regard to the Forms task force there is a quite evolved spec 
out there, even with some implementation in Opera already, and Canvas in 
the other browsers—except IE—too, but I am also very interested in 
hearing more about this ‘XForms Transitional’ idea that was mentioned. I 
have looked at Web Forms 2.0 myself when creating a forms validation 
solution for the Backbase 4.0 AJAX library, and I found the Web Forms 
2.0 stuff to be fairly simple and limited. XForms wasn’t exactly what we 
were looking for either, we weren’t going to implement the whole thing 
(yet), so in the end we ended up having validation based on XML Schema 
Datatypes (like XForms) because they allowed better expression of 
validation rules, data types and inheritance, but in a somewhat simpler 
way (like Web Forms 2.0). I wouldn’t mind elaborating on this, but I 
won’t do it now :).

Similarly, there are several things in XHTML2 that are not in ‘HTML5’ 
but which I think are architecturally more sound and should be in there.

So with the above in mind, and the two suggested approaches, pending any 
better ideas, I think I’d prefer to take the 2-approach for both the 
Forms task force and the HTML group. It will leave more room for some 
fresh ideas and discussion, and integration of various specs instead of 
using HTML5 as a starting point, and at the same time not waste the 
WHATWG’s work.


~Grauw

-- 
Ushiko-san! Kimi wa doushite, Ushiko-san nan da!!
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Laurens Holst, student, university of Utrecht, the Netherlands.
Website: www.grauw.nl. Backbase employee; www.backbase.com.