[whatwg] Questions: IE 6 Compatibility, HTML 5, Spec Timeframe, and Implementation Timeframe

[Ian Hickson]

On Tue, 26 Apr 2005, Brad Neuberg wrote:
> 
> First, what exactly is the stance in regard to IE 6 compatibility for 
> Web Forms 2.0 and Web Applications 1.0?

Basically:

 * New features must gracefully fallback to legacy UAs: although that 
   fallback may be simple lack of support for that feature, using new 
   features in legacy UAs must not cause the experience in older UAs to
   be worse than if the feature was simply not used.

    Examples:

     * <object></object> in HTML4 allows graceful fallback, and is fine.
     
     * <img alt=""> doesn't allow good fallback, but degrades to nothing
       at all, so could be considered if there were no other better ideas.

     * Switching to a different MIME type makes the file unusable in older
       browsers, so it would be unacceptable.

 * Ideally, new features should be implementable using shims in WinIE, but 
   there may be cases where that's not possible, and in those cases we're
   not going to avoid adding the feature just because WinIE can't do it.

    Example:

     * a 3D context for <canvas> is probably not something we can 
       realisticly expect to see implemented in IE using JS, but it's 
       still something we've had demand for and thus something we'll 
       likely be working on.


> I've been hearing things lately concerning Web Applications 1.0 that 
> seem like they would be very difficult, impossible, or cause slow 
> performance if emulated in IE 6.  Whats the exact relationship between 
> these specs and IE 6?  Will there be a baseline of support in IE 6, a 
> low water mark?

The relationship is that most people won't use features that don't work in 
IE, so most features have to bear that in mind. Some people have specific 
needs (<canvas> for example is something we've heard a lot of demand for 
from people wanting to write games and the like), which they can't ever 
expect to really have work in IE, and so for those we need to offer 
features designed so that they can still provide alternative versions for 
IE (i.e. fallback).


> Second, what is the relationship of HTML 5 to these two specs? Who is 
> developing this standard?  At first glance it seems like a large 
> dependency.

HTML5 is the Web Apps spec. It isn't called that yet in the headings for 
political reasons.


> Third, is there a timeframe for completing these two specs and for 
> getting actual implementations out the door?

Web Forms 2 is basically done and will be going to Call For 
Implementations shortly.

Web Apps 1 has no ETA yet. Implementations of some parts have shipped for 
years (XMLHttpRequest), implementations of others are likely to ship soon 
(<canvas>), implementations of other parts aren't likely for a long time 
(relatively speaking).


> I'm concerned that proprietary web app/rich web app defacto standards 
> will succeed faster than the WHAT-WG, like Flash and Avalon, and one of 
> the things that attracted me to the WHAT-WG was its focus on being 
> real-world and pragmatic, getting it out the door rather than getting it 
> perfect, co-opting and using existing de-facto standards like innerHTML 
> rather than rolling new ivory tower ones.  Would hard deadlines on both 
> specs, including deadlines for implementations, help this?

I agree that we have to move fast. I believe the main ways to do this are 
to (a) write text at a steady rate (as I am doing), (b) to get feedback on 
the spec (as is happening), and (c) to stop adding new features. There is 
one more feature I think we need to add to the spec that isn't there 
already, namely the session stuff that people have been discussing. Other 
than that I'm of the opinion that we have enough features for "HTML5" now 
and so "all" that remains is fleshing the spec out.

I don't think deadlines would help, really.

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