W3C home > Mailing lists > Public > www-style@w3.org > March 2012

Re: outline-radius

From: Paul Bakaus <pbakaus@zynga.com>
Date: Tue, 13 Mar 2012 09:30:44 +0000
To: Brian Manthos <brianman@microsoft.com>, Brad Kemper <brad.kemper@gmail.com>
CC: "Kang-Hao (Kenny) Lu" <kennyluck@csail.mit.edu>, Boris Zbarsky <bzbarsky@MIT.EDU>, WWW Style <www-style@w3.org>
Message-ID: <CB84CF89.177D5%pbakaus@zynga.com>
Hi everyone,

Great to see a lively discussion around this topic. A couple of thoughts
regarding the last couple ideas:

- Are there any known instances of developers actually *using* and
preferring the outline wrapping around multiple boxes in situations where
the outline is set via CSS? I just can't figure out why a dev would want
that - it might make total sense as an accessibility thing, but I can't
see it work for layout purposes (If it would wrap around actual
transparent edges - especially on images - that would be different, and we
would totally use it). I am starting to get the feeling that outline is
simply overloaded, and isn't really helpful to address my use-case.

- On first sight, I do like the proposal by Marat regarding multiple
borders and some that affect the box, others that don't. I can see some
tricky situations though with the order of multiple "size affects". I do
think that border would be the logical place to do effects such as mine,
instead of outline.

Yes, I don't know the exact syntax, but getting rid of the outline prop
(or defining it so that it doesn't show an outline per default, but simply
controls the outline that is shown on focus), and extending border seems
like the natural and right thing to do imho.

Thanks,
Paul

Am 13.03.12 02:44 schrieb "Brian Manthos" unter <brianman@microsoft.com>:

>In that case, maybe fully defining the behavior of (rectangular)
>'outline' for spans that split into multiple boxes would be a good start.
>
>Last I checked a lot of outline was underspecified (i.e. lots of UA
>freedom to do all kinds of interesting,
>highly-unlikely-to-be-interoperable things).
>
>It sounds like we would likely be in agreement that the sequence should
>be:
>1. add outline to a L3 module (last I checked CSS2.1 had the latest
>normative text on the property)
>2. clearly specify behavior of "sharp" outline on multi-box spans
>3. clearly specify how border-radius impacts "rounded" outlines
>(including examples from 2)
>4. (L4?) add outline-radius as optional way to directly specify an
>alternative to 3
>
>
>-----Original Message-----
>From: Brad Kemper [mailto:brad.kemper@gmail.com]
>Sent: Monday, March 12, 2012 4:39 PM
>To: Brian Manthos
>Cc: Kang-Hao (Kenny) Lu; Boris Zbarsky; WWW Style
>Subject: Re: outline-radius
>
>
>
>On Mar 12, 2012, at 3:44 PM, Brian Manthos <brianman@microsoft.com> wrote:
>
>> I think you're misunderstanding me here.
>> 
>> Shiny-27 is outline-radius.
>> Shiny-15 is border-radius.
>
>My point is that the way Shiny-15 interacts with 'outline' is broken, and
>should be fixed. We don't need 'outline-radius' to fix it (it should be
>automatic), but as long as we go to the trouble of fixing it, it might be
>nice to have the property too, for manual override or for having rounded
>outlines on sharp-cornered border boxes.
>
>I'm not sure how I feel about negative 'border-radius' lengths to get
>rounded outlines on sharp-cornered border boxes. It feels wrong, and
>would still lead to authors creating almost-but-not-quite-zero border
>radii. 
>
>
>
Received on Tuesday, 13 March 2012 09:31:25 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 22 May 2012 03:48:51 GMT