RE: [css3-images] linear-gradient keywords and angles are opposite

[Brad Kemper:] 
> On Jun 8, 2011, at 4:07 PM, Brian Manthos <brianman@microsoft.com> wrote:
> 
> > From: Brad Kemper [mailto:brad.kemper@gmail.com]
> >> But anyone
> >> using the keyword 'left' knows that it means 'from left to right', or
> >> they find out within seconds and then move on.
> >
> > I disagree completely on this point.  I constantly find this backwards
> notation a stumbling block.   The reason I raised the issue is because I
> don't want to just "move on".  The spec is broken and inconsistent, and I
> thought the goal of having drafts at all was to address such things so
> that the future can be better.
> >
> > Using that argument we might as well call it "foo" because "people will
> just learn that foo means that it progresses from left to right".
> 
> Not so. Currently 'left' has a clearly defined meaning of "start the
> gradient on the left". There is a logic to it that is not difficult to
> learn. It is not a totally arbitrary name.

We know what it is *currently*. The issue is whether it should have this
meaning. That it can be learned is not the question either. I mean, people
can learn floats, too ! 

Whether the current logic does stand up in other use-cases is a question here. 
In particular, I'm interested in making the feature coherent for all users, 
including those who animate its value. It's not at all obvious to me and others
that the current model is coherent in the latter case and that is a concern.

> 
> >
> > Honestly, I think that would be better than using "left" to mean the
> exact opposite of "leftwards".
> 
> I'd be OK with "left" meaning "to the left" instead of "from the left",
> except that means EVERYTHING already authored would produce incorrect
> results in new browsers, and EVERYTHING authored to the new revision will
> produce incorrect results in current versions of browsers that support
> linear-gradient. I don't recall any similar type of situation in CSS with
> a property or value that had already seen enthusiastic author response. So
> I cannot in good conscience support such a change unless linear-gradient
> and/or it's prefixed versions are renamed (e.g. -moz2-linear-gradient or -
> moz-straight-gradient). I'd want that even if I loose the battle to keep
> the more intuitive version of degree direction.

"I cannot in good conscience support such a change unless linear-gradient
 and/or it's prefixed versions are renamed (e.g. -moz2-linear-gradient or -
 moz-straight-gradient)". 

You've just answered your own concern: this is a vendor decision. As are
all decisions pertaining to prefixed *proprietary* property. As a side
note, I will remind you that prefixed properties are invalid per CSS; as 
such it would imo be awkward for the WG to accept/reject a change based on 
proprietary decisions related to invalid CSS.

Received on Thursday, 9 June 2011 00:37:19 UTC