Re: Differences between the W3C and WHATWG specifications

On Tue, 15 Jun 2010, Ian Hickson wrote:
> On Tue, 15 Jun 2010, Maciej Stachowiak wrote:
> 
>> 2) In addition, we have some concerns about the following newly added
>> paragraph:
>> 
>> "The specification published by the WHATWG is not identical to this
>> specification. The main differences are that the WHATWG version includes
>> features not included in this W3C version: some features have been
>> omitted as they are considered part of future revisions of HTML, not
>> HTML5; and other features are omitted because at the W3C they are
>> published as separate specifications. There are also some minor
>> differences. For an exact list of differences, please see the WHATWG
>> specification."
>> 
>> We think this may give the wrong impression about the nature of the
>> differences between the W3C draft and the WHATWG draft. We have the
>> following concerns:
>> 
>> (a) The text implies that all omitted features are either published as
>> separate features, or necessarily part of a future version of HTML.
>> While the WG has not ruled out including removed features in future
>> versions of HTML, we have not committed to doing so, either.
> 
> This working group isn't chartered to work on future versions of HTML, so
> it can't really have an opinion on that. The WHATWG is working on future
> revisions of HTML today. Therefore even if the W3C never published the new
> features, they would be part of "future revisions" of HTML. The text
> therefore seems accurate on this front.


The text in the current draft of W3C specification continues to refer readers to text that implies that all omitted features are either published separately or necessarily part of a future version of HTML. In context, readers are likely to conclude that this means a future W3C version of the specification, even if you intended some other meaning. It's misleading to have text with that implication, whether or not it has some other possible meaning.

The chairs agree that this is unacceptable. Adopting our proposed replacement paragraph would address this issue to our satisfaction.

>> (b) The text implies that changes other than feature removals are minor,
>> but that is debatable. It's better not to make such a judgment.
> 
> I don't see such an implication.

The text in the current draft of W3C specification continues to states "There are also some minor differences," immediately after mentioning "the main differences". Even if you don't see the implication, reasonable readers are likely to conclude that any differences which are not among "the main differences" are "minor differences".

The chairs agree that this is unacceptable. Adopting our proposed replacement paragraph would address this issue to our satisfaction.

>> (c) The WHATWG specification is ever-changing, so we can't be sure at
>> the time of publication of a particular Working Draft that the
>> referenced list of differences will continue to be correct.
> 
> True. I've changed the build process so that the TR snapshots of the spec
> get a different paragraph that is more accurate in that context:
> 
> <p>This version of the W3C HTML5 specification is a snapshot of part
> of the work done by these groups as of [LONGDAY] [YEAR]. Because the
> HTML specification is continuously being maintained, implementors
> and authors are strongly urged to read the latest editor's draft
> instead of this snapshot. The W3C and WHATWG editor's drafts of HTML
> differ in various ways; the reader is referred to <a
> href="http://www.whatwg.org/specs/web-apps/current-work/#is-this-html5?">the
> WHATWG's editor draft</a> for a summary of the current
> differences.</p>

The text continues to refer the reader to the current editor's draft of the WHATWG which does not reflect the list of differences from a W3C perspective.  The chairs agree that this is unacceptable. We request that this link be removed.  Replacing the link with the exact differences from a W3C perspective in the W3C draft could be an acceptable solution, and is something we are willing explore for a future draft. But at this time we are asking for simple removal so we can move forward.


Based on the above, we are requesting that you adopt our previously proposed text as-is rather than debating this at length:

>> "The specification published by the WHATWG is not identical to
>> this specification. The main differences are that the WHATWG
>> version includes features not included in this W3C version: some
>> features have been omitted, but may be considered for future
>> revisions of HTML, beyond HTML5; and other features are omitted
>> because at the W3C they are published as separate
>> specifications."


Regards,
Maciej Stachowiak
(on behalf of all three co-chairs)