[whatwg] Administrivia: Update on the relationship between the WHATWG HTML living standard and the W3C HTML5 specification

[David Bruant]

Le 25/07/2012 13:45, Bronislav Klučka a écrit :
>
> On 20.7.2012 14:38, Steve Faulkner wrote:
>> Hi Hixie,
>>
>> I believe you have made some spurious claims, one of them being;
>>
>> "The WHATWG effort is focused on developing the
>> canonical description of HTML and related technologies"
>> ................
>> The claim that HTML the living standard is canonical appears to imply 
>> that
>> the requirements and advice contained within HTML the living standard is
>> more correct than what is in the HTML5 specification
>> ................
>> In respect to those author related requirements mentioned above the 
>> HTML5
>> specification can currently claim to be contain a more accurate set of
>> requirements and advice, that takes into account current implementation
>> realities, thus providing author with more practical advice and thus end
>> users with a better experience.
>>
>>
> Canonical means neither "correct" nor "accurate", those words have no 
> meaning in this case, you cannot apply them on set of rules (you first 
> have to have set of rules, to claim, whether something is accurate or 
> correct within the boundaries of those rules), canonical means, that 
> those set of rules are valid, that those rules apply.
> The question is, who will follow those set of rules. Both HTML5 and 
> HTML TLS can claim to be canonical, both can be valid for different 
> groups.
Who are those different groups? As far as I know, the major producer of 
HTML content is a group usually refered as "web developers".
I have a genuine question which is: does what any other group think 
about HTML matters?

> Let's just hope all major vendors will chose the same...
This statement seems very upside down to me. Some people thinks 
standards follow this path:
1) a spec is written
2) some folks (browser vendors) implement it
3) some other folks (web devs) build on top of that.
I don't know how to say that, but this model is plain wrong. Maybe it 
works in some industry. Maybe it works for some languages, but it's not 
the case for web technologies (HTML included).

Reality is that implementors implement and for the most part, the spec 
codifies *afterwards* what has been implemented. The biggest part of 
HTML5 has been exactly that. And sometimes between the first 
implementation and the standard, other implement the same thing. 
Sometimes with bug, sometimes they get the initial implementors to fix a 
bug before copying it, sometimes no one care about what has been 
implemented and it's removed.

In the end, all majors vendors will keep doing what they have always 
been doing and for most part, since they all have the constraint to 
properly render at least what others render properly, there will be at 
worst a de facto standard if the standard either doesn't cover this case 
or is not properly describing what is actually implemented.

Who writes the spec, what is in the spec is at no point the most 
important piece of the puzzle, because if no one follows the spec, it's 
just a useless piece of document.
My good news for you is the following: all relevant specs will codify 
what all major vendors choose to do.

David