Re: [css3-background] border-radius color transitions using gradients recommended but undefined

On 02/23/2010 02:52 PM, Sylvain Galineau wrote:
>
>> I just told you, I can't work with this.
> I heard you. Did you hear that we can't work with this either ?

Yes, but you aren't telling me what you *can* work with.

>> How about you stop defending yourself and answer the goddamn question?
>
> I'm attempting to defend authors, actually. They have enough interop pain to
> deal with already without being opted into an experiment of each browser's
> definition of border gradient prettiness.

I think authors would much rather have minor differences in border gradient
prettiness than muck about with sliding doors techniques for the next four
years. There's different levels of interop pain, and ones that involve layout
differences and brokenness are more painful than those that trigger minor
differences in rendering when none of those strike the target audience as
noticeably "wrong".

> The answer Elika is that I don't know. That's the whole "goddamn" point. I don't
> know anymore than you do and I've been looking at bloody gradients until I cried.

Ok, I got that.

> I don't have any more psychic abilities than you do and have no idea what authors
> will want. I'm not even sure everyone on this mailing agrees what should be done.

I'm pretty sure I can figure out a really good answer to "What's the best way to
draw a gradient for a solid-solid join?" As I've said many times before, I'm happy
to work with your developers to design an implementation. The question *here*
is, to what level of detail do you need it specified *in the W3C spec*?

You don't need to tell me the gradient arguments for me to write the spec -- I
can figure those out myself. But I need to know the parameters you want defined
in the spec, because the WG told me "color stops and gradient shape", and now
you're telling me that's not sufficient. (Well, actually, you're refusing to
answer my question of whether that's sufficient.)

> I'll thus suggest again - and once again you will conveniently ignore it - that this
> entire discussion has simply established that this behavior is not suitable to be
> recommended in this level of the spec. And deserves experimentation on an opt-in basis.

If the opt-in means requiring a sharp color join in CSS3 and enabling gradients
via another property in CSS4, then that's not acceptable to me.

~fantasai

Received on Wednesday, 24 February 2010 00:02:32 UTC