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

Re: % is not a valid unit for border-width

From: Matthew Wilcox <elvendil@gmail.com>
Date: Fri, 20 Jan 2012 13:13:07 +0000
Message-ID: <CAMCRKiJRzA_to3rWVGeyM_xMakpfP=i4ur+3y1DswMc8gCxqaQ@mail.gmail.com>
To: "L. David Baron" <dbaron@dbaron.org>
Cc: Lea Verou <leaverou@gmail.com>, www-style@w3.org
A recent article on "the bane of responsive layouts" - a lack of % for

The solutions are poor - either you use outline and get errors with Firefox
(and poor line placement) or you use box-sizing, but box-sizing can't
degrade gracefully.

We could do with % for border-width.

This inability to gracefully degrade is also why I'd like to see
conditionals. When you change box-model you have to change other values -
but those values have no idea which box-model they're working inside of.

On 6 January 2012 10:10, Matthew Wilcox <elvendil@gmail.com> wrote:

> Sorry Tab, typo in the url, I did indeed mean http://adaptive-images.com
> I'm glad that's convinced you :)
> @ David
> I can see why that's an issue for you. I would put forward that when
> %units are used the priority changes - it's no longer about ensuring
> the border is precisely rounded to give an even look, but about
> accuracy of layout. I'd also agree that it would make sense for CSS to
> specify how that's handled.
> To be clear though, this is a technique that is already in us,
> becoming mainstream, and will be used more and more given how mobile
> devices are set to outstrip desktop by 2015 (and it is in that context
> where it matters most, at the moment). Because it is used for layout
> it is vital that it is 'fixed'.
> Additionally, with higher resolution devices becoming increasingly
> popular (like retina displays) the problem of "uneven" borders will
> become less and less noticeable. It's only an artefact of having a
> limited number of pixels on the display, and that's becoming less of a
> limit as time wears on. Thinking of the future, I'd like to suggest we
> downplay the limitations of current displays and go with sacrificing
> border width consistency around the box instead of some other option.
Received on Friday, 20 January 2012 13:13:37 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:38:54 UTC