layout solutions)ions of CSS for years to come (was RE: CSS multicolumn layout solutions)

Tantek Çelik observed:

> Neither HTML4 nor CSS2 would have exited "CR"
> had there been a CR period at the time they
> were made "recommendations".

You have my full agreement.

> IMHO W3C "Recommendations" issued before the
> CR period was established should have been
> immediately reverted to CR status in order
> to pass the updated requirements of two or
> more interoperable implementations.  In lieu
> of this, I personally consider those W3C docs
> to be in CR until such time that strict CR exit
> criteria have been met.

I sympathize, but, if I may veer off topic, I question the feasibility 
of reverting a final Recommendation to the status of 
Candidate Recommendation.  This is equivalent to stating, 
"Yeah, that specification that we publicized and promoted and 
told you to adopt--well, it might not be such hot stuff after all.  
Give us a while to work out the kinks and *then* we can give 
you something that you should use."

> IE5/Mac was released almost two years
> ago with fully conformant HTML4 and CSS1,
> and IE6/Windows was released in the past year
> as well with fully conformant CSS1 and DOM1.

And I scrambled to obtain both as soon as they hit the Web.  
How many people literally wept with joy upon seeing the 
powerful CSS implementation in Internet Explorer 5 
Macintosh Edition?  I did, Tantek.  The initial euphoria 
eventually passed and then I got down to business: hunting bugs.

I am sorry to bear this message, but neither Internet Explorer 
5 Macintosh Edition nor Internet Explorer 6 
conformantly implements CSS1.  Not counting the 
CSS1 conformance violations that result from implementing parts 
of CSS2 and CSS3, bugs remain.  Some of the bugs are trivial; 
others are major.  (Internet Explorer 5 Macintosh Edition also 
has HTML4 bugs, but that discussion belongs in another forum. 
 Lacking testing, I decline to comment on the DOM1 
implementation of Internet Explorer 6.)

> I know it may be far more enjoyable to stay
> "irritated"

The benefit of my pessimism is that I will either be correct or 
be happily surprised.

>> CSS, in particular, is an area that vendors will be slow to
>> implement conformantly.
> Again - there is that memory loss of the past
> two years or so.  I count at least five
> reasonably solid CSS implementations on a
> plethora of platforms:
>  IE5/Mac
>  IE6/Windows
>  Opera5,6
>  NN6.x
>  Konqueror (haven't verified, but have heard
> good things about it from folks I respect)

I wrote "conformantly" for a reason.  I readily agree that 
the applications that you listed have "reasonably solid 
CSS implementations".  I would, in fact, go further: 
those applications have truly good CSS implementations.  
Yet neither "reasonably solid" nor "truly good" is synonymous 
with "conformant".  Of all people, we on this list should be able 
to make that distinction.

Conformance is not the soul and lifeblood of Web software.  
Most people, myself included, value speed, stability, 
and convenience as much as if not more than they 
value conformance to specifications.  This provides no 
excuse, though, to dilute the meaning of conformance.

Etan Wexler

Received on Friday, 25 January 2002 08:26:31 UTC