[whatwg] Switching to an unversioned development model

[Ian Hickson]

On Sat, 9 Jan 2010, Philip Jägenstedt wrote:
> 
> Is WHATWG HTML is intended to stay at "Working Draft" indefinitely, or 
> does the spec status not really mean anything at this point?

The spec status is given by the little notes in the margins, which give 
the state of each section. This is unchanged; we've been using this method 
for over a year now. Anyone should feel free to keep those markers updated 
if they notice they are getting out of sync with reality.


> > For people who were paying _really_ close attention, the separate 
> > WHATWG Microdata Vocabularies draft is also gone (the vocabularies are 
> > now defined in the WHATWG HTML spec itself again). I got rid of them 
> > as part of a simplification of the specs being generated, and because 
> > the W3C's equivalent specs were all merged into a single HTML5 
> > Microdata subspec.
> 
> Yes, I was wondering if this was a bit of an accident or what. Is it the 
> intention to keep these in the WHATWG HTML spec indefinitely and even 
> add more vocabularies as they become needed/popular?

I expect we'll figure this out as Microdata gets deployed -- it may be 
that Microdata gets deployed widely but that nobody really uses standard 
vocabularies, such that we don't need them at all, or it may be that 
standard vocabularies are the main way it gets used, in which case it 
makes sense to have them right there with the microdata spec, or it may be 
that they are used occasionally but only as a small part of a wider set of 
vocabularies, in which case authors would benefit more from them being 
independent of the main spec, etc. I guess we'll play this by ear.


On Sat, 9 Jan 2010, David Bruant wrote:
> 
> As said in the whatwg wiki "The point to all this is that you shouldn’t
> place too much weight on the status of the specification as a whole. You need
> to consider the stability and maturity level of each section individually. ".

Indeed.


> However, I think I understand the underlying idea which is that if we keep
> adding things, the HTML5 spec will never reach the recommendation status (even
> for 2022).

At this point, the snapshot that is HTML_5_ is being driven through the 
more formal "WD", "LC", "CR", "REC" status by the W3C, and excludes the 
newer additions (specifically currently <device>).


> The <device> element offers a very interesting potential. However, I 
> have the impression that it will require years of work before having a 
> first good "working draft" because of diversity of devices sorts and 
> even devices serving the same purpose.

Actually right now the main blocker is the lack of a good common video 
codec (that old chestnut again!). There's not much point having it without 
a common codec -- if it's unlikely that static video providers will 
provide two or more video formats, it's even more unlikely that sites will 
be ready to do on-the-fly transcoding for video-conferencing.


> In my humble opinion, it would be wiser if the device element had its 
> own separate specification and wasn't lost within an already almost one 
> million pixels height spec.

If <device> turns out to be a good idea that works, then I think it will 
benefit from being defined along with all the other parts of HTML. When 
writing Web Forms 2, we found that it was really inconvenient for everyone 
for content to be split across multiple independent specs.


> While reading the whatwg wiki section "When will HTML5 be finished? ", I 
> realized that it is currently hard for web authors to answer the 
> question "what sections of the spec are at the 'Implemented and widely 
> deployed' status ?" It might be a good idea for communication around the 
> HTML5 spec to keep an updated list of sections and subsections sorted by 
> level of maturity. Another (complementary ?) idea would be to provide 
> this information in the table of content by putting titles of 
> sections/subsections in different colors (and/or background color). It 
> would be a good way for people who are discovering HTML5 to understand 
> what they are somehow "allowed" to use today.

Agreed. If anyone wants to volunteer to either maintain such a list or 
make changes to the status.js script to annotate the spec in such a way, I 
would be more than happy to help out. Please let me know!

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'