Re: [css3-images] Proposed Gradients changes

On Nov 18, 2010, at 4:01 PM, "Tab Atkins Jr." <jackalmage@gmail.com> wrote:

> On Thu, Nov 18, 2010 at 2:11 PM, Brad Kemper <brad.kemper@gmail.com> wrote:
>> On Nov 18, 2010, at 10:22 AM, "Tab Atkins Jr." <jackalmage@gmail.com> wrote:
>> And then the computed value of the direction would always be an angle.
> 
> Hrm, for some reason I didn't think of it in that terms.  That should
> work just about as easily; rather than saying all the gradients are
> two-point underneath, they're all angle underneath.

Exactly.  

> I don't like the keywords you have, though - I'd rather just restrict
> it to "[ [left | right] || [top | bottom] ]" and still parse it like a
> background position (that is, "left bottom" and "bottom left" both
> work and mean the same thing).

I won't put quite as much energy into debating that, as we are now much closer to agreement than before. :) 

I do still prefer one keyword rather than 3-5. Or even three and always three ('bottom-left to top-right' and 'left to right'). One keyword works best for me, because there are only eight variations and direction is then as simple as "pick an angle or a keyword", and the canonical way is the commonest English-language way, with no confusing this with actual bg-position syntax. Border-radius has that sort of thing as part of its longhand property names for describing the same sort of corners on element boxes ('border-radius-top-left', for example). Before I had 'border-radius' syntax memorized (and when I was used to FF doing it differently that Safari), the hardest part to remembering if the corner name came before or after the "radius" part of the property name. I never had any problem remembering the order of "top-right" and "bottom-right", etc. Because that is the way people say them naturally (unlike with 'background-position', where I STILL have problems remembering the proper order rules, which you saw in my earlier email).


> This does make something that is strictly weaker than what exist
> currently, but I think it still hits the 90% mark,

Probably even more than that. My 'as-square' keyword was meant to pick up most of the remainder (for other angles that could stretch with the box), but yeah. This could be good enough and keep the simplicity. 

> and the
> interpolation is somewhat prettier.  And, like I wanted, it's still
> compatible with most gradients in practice.

Exactly. We can wait and see if we could pick up another 5% or so with a 'as-square' or 'stretchy' or whatever keyword, and if that's important enough for the extra "indirection".

Received on Friday, 19 November 2010 02:01:51 UTC