W3C home > Mailing lists > Public > www-style@w3.org > December 2010

Re: [css3-images] Repeating oblique gradients

From: Tab Atkins Jr. <jackalmage@gmail.com>
Date: Wed, 1 Dec 2010 10:22:15 -0800
Message-ID: <AANLkTikxcVA=7xrM0DhR+JRfwnJJ+V5812wfJfa2Eoft@mail.gmail.com>
To: Brad Kemper <brad.kemper@gmail.com>
Cc: Leif Arne Storset <lstorset@opera.com>, Rik Cabanier <cabanier@adobe.com>, www-style list <www-style@w3.org>
On Tue, Nov 30, 2010 at 8:48 PM, Brad Kemper <brad.kemper@gmail.com> wrote:
> What was unclear in my last e-mail?
> You mentioned 'background-repeat:extend' as though it was a done deal, and I
> reminded you that we had also discussed making the extend behavior automatic
> for any non-repeating "background-repeat" value (so no "extend" value is
> needed). I don't have a strong opinion either way, but was laying out the
> differences.

Ah, ok.  No argument there, then.

> You said resizing the image is irrelevant to the discussion about filling
> the background (or perhaps about tiling too, as I had talked about in the
> part you quoted). I defended why I had brought it up: because it lets you
> see several tiles at once (if tiling), or the area outside the image (if
> not). Thus, you can't just pretend that tiling gradient images in
> backgrounds won't exist. When combined with 'background-size', it is a
> rather obvious way to get repeating gradients.

I claim it's irrelevant because, well, it is.  ^_^  Gradients are
finite-sized regardless of the value you give to background-size, and
thus the interaction between gradients and tiling/etc is always
present.  You seemed to be stating that setting background-size to a
non-default value changes things somehow, which is incorrect.

> You said it was a "happy coincidence" that tiling gradient images can
> produce nice effects of repeating patterns. I responded that nice effects of
> repeating patterns is exactly what background-repeat is for. I'd say it was
> a more of a "happy coincidence" that you can do something similar inside the
> image by adding more and more repeating color stops, or by adding a keyword
> to repeat the color stops, except that it is really rather deliberate to
> have the ability to do so. I think it is absurd for anyone to take the fact
> that repeating gradient images as tiles in a background mostly works well
> (but needs some improvement with diagonals), and say it has no bearing on
> the "general problem of a repeating gradient". The gradients you are working
> on is in an IMAGES module, so a repeating image and a repeating gradient DO
> have bearing on each other, because the gradient IS an image. Since authors
> will be repeating these images in backgrounds, why would you NOT want them
> to work well when the gradient direction is angled, now that we have the
> opportunity to do so?

Correct, gradients are images, and you tile images.  My point is that
tiling images is a *different thing* than repeating gradients; the two
just happen to produce the same result in a few limited circumstances.
 Tiling treats the image as a block to be stamped out in an infinite
pattern.  Repeating gradients instead modifies the base image itself
so that it has infinite color-stops in a regular pattern.  It is not a
repetition of the original image, but rather a modification.

It could be that we're just abstracting the ability in different ways.
 I consider the tiling effect as I state above.  You appear to
consider it as a more general effect that is specialized to the image
type.  I don't fully understand your position, so I may be misstating.

Received on Wednesday, 1 December 2010 18:23:08 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:34 GMT