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

fantasai:
> Firstly, I think it would make more sense to ask the authoring
> community
> rather than the implementing community whether having a "left" gradient
> start at the left side of the box is confusing. They, not you, are the
> ones using the syntax, and they, unlike you, aren't thinking about the
> internal model of the implementation but about consistency with how
> they
> would themselves describe their intentended result.
 
My concern has nothing to do with internal implementation.  The implementation isn't easier or harder either way.

My concern was both from the authoring side of gradients alone and from the transition aspect that the WG apparently hasn't fully thought through yet.

Don't shoot the messenger because your mental model aligns with an inconsistent system and you're willing to accept the inconsistency because you've grown accustomed to it.

> Secondly, I don't agree that the old spec is in any way inconsistent on
> how it defines the named keywords.

Just because you don't see the inconsistency, doesn't mean it isn't there.  I've pointed out the literal juxtaposition of the text in the spec in multiple places, multiple times, in multiple version of the spec which shows clearly the inconsistency.

Tab didn't see it initial, but now does.  Alan didn't see to see it until yesterday.


Simon appears comfortable with the inconsistency.

>From my perspective, a loose paraphrasing of Simon and your point goes something like this...

"When I see foo(M unit a,b,c) and foo(N unit b,b,c), it doesn't matter if the two uses of the functional syntax use a completely different model for the two rendering as long as 'a' and 'b' are different."


To me that's accepting inconsistency, to you it's not.  Apparently, we'll never agree on this point and that's sometimes the way of it.


But please don't bend off into tangents that assert (incorrect) motives for raising the issue in an attempt to support your position.

Again, this has nothing to do with implementation cost.  It has to do with screwing up the syntax and having to fix it later once the WG recognizes the mistake in a CSS.[next]+ draft.

Received on Wednesday, 15 June 2011 17:18:39 UTC