RE: [HDP] Missing principles

> From: Karl Dubost [mailto:karl@w3.org] 
> Sent: 24 August 2007 01:08
...
> > [1] Deliver incrementally (or similar words).
> > [2] Collaborate with implementers (or similar words).
> 
> Are these Technology design principles or charter principles?

I think both, like numerous others of the current principles.


> 
> This WG is chartered only for one version of HTML.
> Or do you mean the WG should publish, let's say, HTML 5 with 
> only the parser rules and nothing else (no added semantics).

The content of the 'base' version would need discussion.  The key point is to produce something within a year or so, then regularly improve that according to a prioritisation of needs, rather than keep people waiting for a huge change in 10 years time, without the guarrantee that the whole spec will be supported by most user agents.

> then next revision, 3 months later, adding video element  → 
> HTML 5.1 then next revision, 3 months later, adding canvas 
> element → HTML 5.2
> 
> It could be an interesting idea. Though it makes the CR phase 
> a bit complicated.

I'm not sure why, unless the issue is resources.  As long as things progress in order.


> 
> 
> The point 2. is already the case And something which is (or should
> be) the exact nature of W3C.

I think it should be the case, though perhaps not in exactly the same way for work on completely new technologies as for work on things like HTML and CSS.  I think HTML and CSS are now in a different kind of mode than the initial big bang of new technology development, and with so many users interoperability becomes very important.  Interoperability from a user perspective comes not only from standards, but from being able to rely on a feature being supported, and supported in the same way, in a wide range of user agents.

What I'm suggesting, however, is more than just general cooperation and discussion, and probably requires the incremental delivery approach to produce a manageable set of work items.  

I'm suggesting that we plan which features will go into, say, HTML 5.1 by first securing agreement from a 'quorum' of implementers that they will actually implement the proposed features.  If, say, Mozilla, Safari and IE won't agree to implement a particular feature at this time, then we don't specify it for that release.  Of course, Mozilla, Safari and IE would need to be take the needs of users/authors into account when making these decisions.  The standardisation work would then benefit immediately from implementation lessons-learned and probably test data. In this way, standardisation and implementation efforts advance together, leading to true interoperability for users.

One reason I don't think this is already the case, is that I'm not aware of any such explicit commitments among implementers, users and spec developers that would describe the scope of the current effort. 

What I'm suggesting is perhaps a little like the CSS 2.1 effort, but NOT as a retrofit.  Rather as an ongoing, forward-moving way of working.  So perhaps a good starting point would be to produce an initial version of the HTML5 spec that reflects mostly what is largely supported in current user agents, but addressing some of the gaps and ambiguities etc in the HTML 4.01 spec that currently cause problems.  After that issue new versions that are incrementally carved out of plans for addition of new features or resolutions to current interoperability issues.

RI

============
Richard Ishida
Internationalization Lead
W3C (World Wide Web Consortium)
 
http://www.w3.org/People/Ishida/
http://www.w3.org/International/
http://people.w3.org/rishida/blog/
http://www.flickr.com/photos/ishida/
 
 

> 
> --
> Karl Dubost - http://www.w3.org/People/karl/ W3C Conformance 
> Manager, QA Activity Lead
>    QA Weblog - http://www.w3.org/QA/
>       *** Be Strict To Be Cool ***
> 
> 
>

Received on Friday, 24 August 2007 12:52:48 UTC