[whatwg] Tim BL's HTML WG announcement and WHAT WG

[Charles McCathieNevile]

On Mon, 30 Oct 2006 14:57:39 +1000, Lachlan Hunt  
<lachlan.hunt at lachy.id.au> wrote:

> The fact is that whatever we define must be compatible with desktop  
> browsers above everything else.

Actually (as the representative of a desktop browser at W3C) I think this  
overstates the case.

>  That is what the vast majority of  authors write HTML for,

If that is the measure, the vast majority of content is written for IE.  
Some of it also gets tested in Firefox, and a small amount gets properly  
tested for desktop.

> it's what they test in and it's how they expect all other tools to  
> handle it.  I'm not saying that getting implementation experience from  
> other tools wouldn't be valuable, just that it's far more practical for  
> other tools to base their implementations on desktop browsers, than the  
> other way around.

In a number of practical cases desktop browsers are based on other tools -  
especially on authoring tools and whatever junk they happen to produce. It  
turns out to be more practical for us, although it is destructive in the  
long run.

More importantly, mobile web browsing is growing - especially in places  
like Korea (where there is virtually nothing except IE on the desktop) and  
Japan, but also in Europe. (The US lags well behind in most mobile  
services because people who use public transport in other countries are  
busy driving, among other sociological differences). And copying things  
 from desktop to mobile is extremely difficult. Microsoft didn't - they  
have a seperate team who are learning to build a mobile browser. Mozilla  
haven't really got a mobile browser. Nokia have taken all their experience  
working with us, and with WAP browsers, plus an existing browser engine,  
and a lot of work, to come up with something reasonable. Meanwhile  
Jaatayu, Infraware, Openwave, Access, Obigo and others who are unknown in  
the desktop world are also producing browsers in this space.

The same goes to some extent for embedding in devices that are not really  
phones but are not really desktops either. A decade of working with TV as  
a Web platform shows that is different too, and that simply copying the  
desktop experience to it is a failure.

These are IMHO important caveats that should be borne clearly in mind when  
working out the place of "what desktop browsers do, and will do" in  
standards development. That said, the role of desktop browsers is clearly  
important...

> Disgusted of Tunbridge Wells wrote:
>> In addition, if the role of the W3C is simply to to place
>> its seal of approval on what /is/, rather than on what
>> /should be/, then I for one would find that a very
>> disturbing (not to say depressing) state of affairs.
>
> The W3C needs to release specs that can actually be useful in the real  
> world, not ones that are just nice in theory, but totally impractical to  
> implement.

Having standards that are practical to implement (an important part of  
this is that they are released as "finished" and we can rely on them not  
being changed) is not the same as having standards that exactly specify  
every last corner case. In any event, expecting authoring tools, user  
agents, hand authors, to all follow standards perfectly within the useful  
lifespan of a particular web standard seems optimistic to me - a goal  
worth striving for, but not necessarily the only consideration for  
measuring success.

> For HTML, if the W3C released another specification that required SGML  
> parsing, it would be completely useless

The odds that W3C will insist on SGML parsing seem to me infinitesimal.  
(Releasing a non-XML version of XHTML 2.0 and calling it HTML 5 would be  
useless too, but I don't see that happening either...)

> because it is practically impossible for browser vendors to conform to  
> such requirements without breaking a significant portion of existing web  
> sites.

It is difficult for us to conform strictly to any reasonable* standards  
without breaking a significant portion of existing web sites. The fact is  
that sites like Google, Microsoft or Yahoo ignore standards seemingly at  
will, and can continue to do so, because browsers will be changed to make  
them work.

*reasonable, of course, means what *I* think is reasonable ;) As simple as  
possible, clear, accurate, supports freedom, believes in truth and  
justice, is wise, ...

cheers

Chaals

-- 
   Charles McCathieNevile, Opera Software: Standards Group
   hablo español  -  je parle français  -  jeg lærer norsk
chaals at opera.com          Try Opera 9 now! http://opera.com