W3C home > Mailing lists > Public > www-style@w3.org > May 2012

Re: [css-variables] Using $foo as the syntax for variables

From: Tab Atkins Jr. <jackalmage@gmail.com>
Date: Wed, 23 May 2012 10:35:39 -0700
Message-ID: <CAAWBYDCdWJEgagbON6QY6qNAhTF6wwt9UXQjJZbZqjqqCKVfHA@mail.gmail.com>
To: Bjoern Hoehrmann <derhoermi@gmx.net>
Cc: www-style list <www-style@w3.org>
On Wed, May 23, 2012 at 10:15 AM, Bjoern Hoehrmann <derhoermi@gmx.net> wrote:
> * Tab Atkins Jr. wrote:
>>If you read my email, I agree with you.  Breaking a non-trivial amount
>>of public content is a strong argument against making a change,
>>because it hurts authors and users (and then, by extension, hurts
>>browsers too).
>>The fact that an impl has to change isn't a strong argument - that has
>>to happen every time we make any change or addition to the language at
>>all.  If a change is *difficult*, that's a moderate argument against a
>>feature.  What I'm suggesting here definitely is not.
> When I talk about breaking style sheets I mean breaking style sheets in
> a way that is notable, not about a couple of browser vendors refusing to
> make a change because it would break too much content they care about.

I have no idea how you define "notable" besides "breaks a non-trivial
amount of public content".

> Again, the problem is that you are creating uncertainty about how CSS
> syntax will evolve, one day you want to disallow comments in numbers,
> next you wonder about comments in variable references, and as there are
> many other places where you might want to disallow comments, I would
> rather you would start one thread discussing what we should expect in
> this area. Would you some day come and say ´p/**/:/**/:/**/before` has
> to become an invalid selector? And would the CSS Working Group always
> agree that wanting to disallow comments is a good enough reason to mo-
> dify the never-changing core syntax? Or maybe it does not want to move
> in that direction, and hence decide that it was a mistake to disallow
> comments between sign and digits in numbers? So, how about starting
> such a thread?

Slippery slope arguments aren't very convincing.  They are also
centered around either theoretical purity or implementor convenience,
which are low on the order of constituencies, so making long-term
binding decisions based on them is a bad idea.

Received on Wednesday, 23 May 2012 17:36:30 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:08:16 UTC