Re: [CSS3, border-radius] and now is real fun.

fantasai wrote:
> [Forgot to CC the list, sorry Andrew for sending twice.]
> 
fantasai wrote:
 > Andrew Fedoniouk wrote:
 >> Following figure
 >>
 >>http://www.terrainformatica.com/w3/border-radius-transition-styles-fig.png
 >>
 >> shows two possible ways (of many) of doing
 >> transition on rounded corner between, say, border-style:inset and
 >> border-style:outset.
 >>
 >> What is better and ideologically correct? Any other ideas?
 >
 > I think the main problem with A is that it introduces transparent
 > gaps in the border for a style that, under normal circumstances,
 > is totally opaque.

Gaps are not the problem, at least they are not worse than in transition
from, say, dashed to double line.

Alfons Mucha [1] and other people from Belle Époque would do it that way.

 >
 > For inset/outset transitions, since they're based on a 3D concept,
 > I think one should consider how that'd look if it were carved out
 > of metal.
 >
 > Other than that, I have no opinion here. I've seen several options
 > that are all very interesting.

But in any case we should define area where transition is allowed to happen.

Or better to define areas of the border where transition does not
definitely occur.

My proposal is simple: area where transition on the corner
is allowed to occur is a rectangle having dimensions max(Rx,Bx), max(Ry,By).

This simple rule covers rounded and sharp corners. It covers also
your idea with the sector that is not always of 90 degree.
But it would be interesting to see formula determining position of the
center of that sector. It appears that location of the center is some 
non-monotone function and that is highly not desirable.
Ideally if you change say X radius value then Y position of the center
should not jump anywhere. Otherwise borders or radius given in
% units will produce sudden changes of rendering on small changes
of size of the element.

 >
 >> Spec [1] is timidly silent about this so asking.
 >>
 >> [1] 
http://www.w3.org/TR/2008/WD-css3-background-20080910/#the-border-radius
 >
 > Spec is silent because the spec writer thinks it's better for the
 > implementors to experiment than to standardize here. :) There's no
 > obviously right answer to many of these transitions.

Thanks, I wish you calm nights beside your fireplace too :)

 >
 > ~fantasai
 >

[1] http://en.wikipedia.org/wiki/Alfons_Mucha

-- 
Andrew Fedoniouk.

http://terrainformatica.com

Received on Wednesday, 26 August 2009 06:03:23 UTC