Re: CSS priorities

On 20/01/2014 04:18, Jens O. Meiert wrote:

> Tab, Working Group members, could you comment on focus and what plans
> you have with respect to curbing feature creep and spec fragmentation?
>
> I had responded with the following to Daniel’s transitions email, but
> predictably that wasn’t very useful there:
>
> […] my point here, or wish, is that it would be good to focus CSS more.
>
> I’ve brought this up a few times in the past but I really think we
> don’t do anyone a favor cramming, pardon a bit of judgment,
> *everything* into CSS. Variables/constants are a great example: They
> should not have become part of CSS, and that not just because authors
> never optimized their CSS well enough to discover that they don’t need
> variables to fight complexity, but simply because anything like PHP
> would do for this [1].

Hi Jens,

You already said it and I already replied. We did not do Variables, we
did something the perfect CSS way. What the authors of CSS Custom
Properties did is a smart, non-intrusive, not competing with
preprocessors, way of adding a bit of intra- *and* inter-stylesheet
salt. Web authors have been constantly asking for such a mechanism
since 1998. In all conferences I have spoken to about the progress of
CSS, CSS Custom Properties have _always_ generated a loud "wow" from
the audience. You can't beat that, Jens. _You_ don't like it, and that's
absolutely fine. FWIW, Bert does not like it either. But this feature
was decided based on a large consensus in the Working Group based on
users' requests. This is the way Standardization works.

> Complexity then is hard to quantify but I think CSS complexity has

CSS Custom Properties are not complex, they're on the contrary
incredibly simple, must simpler than the concept of Variables. If you
want to explain why modularization and feature extension is, in your
opinion, bad, please take another example? Again, the CSS WG did _not_
agree on Variables.

> just exploded because anything that’s asked for and “sounds good” is
> added. I’m getting a little cynical maybe but I don’t recall any
> feature being *rejected* here for about ten years (with the exception
> of obvious nonsense). Now we have 300 properties (up from 53 in CSS 1
> [2]) and are about to make CSS a programming language.

No, we're not. We almost did it, when we discussed Variables. Then
we eventually found a better way to do it. All the provisions we now
have in CSS are stylistic. Most of them were already implemented in
style languages like the P language or DSSSL eons ago. The new ones
address new requests that the whole community in general and the
CSS WG Members in particular feels "in scope". And we often disagree.
Some Drafts we are working on trigger intense discussions, sometimes
intense disagreements. We compromise, we resolve issues to please
everyone, we reach consensus. If we don't, Members (Members of the
whole W3C, not Members of CSSWG only) vote "no" to the REC. That's
as simple.

CSS 1 was a very simple document that was not implemented fully until
MacIE did it, years after publication. CSS 2.0 was done in two years,
untested too; it took us a decade to turn it into CSS 2.1 with a major
cleanup and a full Test Suite. The Modules we published after CSS 3
were started based on both Users and Vendors feedback; we never added
things because they sounded good to be added. As Liam said in another
message, we rejected many, many proposals based on both technical
reasons and use cases. In fact, we have rejected Variables during ten
long years; and we eventually dropped them in favor of Custom
Properties.

Yes, CSS has many more properties than ten years ago. But the Web is not
the same either. Web technologies have reached unexpected fields and
now require features that were impossible to implement ten years ago.
Technology matures, use cases appear, users (hear web authors) have
new requests, and we have to address them the best we can. Sometimes
it belongs to CSS and we put it on the WG's radar *at least to discuss
it*; sometimes it does not and we say it clearly and forward the request
to other WGs or even other Standard Bodies. www-style is full of
requests we refused to address.

Let me come back to your own words: "focus CSS more". Focus more on
what? On Web pages? The times CSS was used only for web pages is faaar
behind us. On desktop browsers? Mobile access to the Web changed that.
On a "notepadable" language? People don't care any more, the times web
authors were all hardcoding their CSS by hand is also behind us. On
a "pure CSS" spirit? Come on, I don't know a single web agency that
never uses a preprocessor for complex web sites.
We _do_ focus on the future of CSS. In the recent years, we made CSS
reach new fields like the Ebooks industry, the automotive industry,
the web-based apps domain and more.

There is one point I can agree with you on: we have to deal with
complexity and trust me on that please, we do. Whenever a feature, old
or new, is discussed, there is always someone to say it loud and clear
if we diverge on complexity and/or readability. Fortunately, we have
super-smart people in the CSS WG, people able to find simple declarative
ways to express complicated stuff. Bert for instance is a true master at
that. Trust us, we do care.

Hope that helps.

</Daniel>
--
W3C CSS Working Group, Co-chair

Received on Wednesday, 22 January 2014 09:01:08 UTC