W3C home > Mailing lists > Public > www-style@w3.org > February 2009

Re: [css21] computed value of top/bottom for relative positioning

From: Giovanni Campagna <scampa.giovanni@gmail.com>
Date: Thu, 12 Feb 2009 22:01:38 +0100
Message-ID: <65307430902121301g44fce838p1f1347099505d36c@mail.gmail.com>
To: www-style@w3.org
Comparing to the general definition of "computed value": "Computing a value
never requires the user agent to render the document" (CSS21, sect 6.1.2), I
agree that the Relative Positioning section should say:

If both the specified values for left and right are not auto, if direction
is ltr, then right is treated as auto, else left is treated as auto. If both
the specified values for top and bottom are not auto, then bottom is treated
as auto. In that case the element is said to be over-constrained.
If both the specified values for left and right are auto, their computed
values are the same as the specified values. The same applies for top and
If the specified value is not auto the computed value is the same as if
position was not relative nor static (absolute length or percentage).
If instead the specified value is auto the computed value is minus the
computed value of the corresponding property.

What do you think?


@Anton: I meant height, I wrote width

2009/2/12 Boris Zbarsky <bzbarsky@mit.edu>

> Anton Prowse wrote:
>> Nice puzzle!  I don't think you're missing anything, and I suppose it
>> depends on what the philosophy of computed values is.  Do specified
>> percentages normally remain as percentages for the purposes of the computed
>> value?
> Yes.
> I think Anne's point is that section 9.4.3 should explicitly say that the
> computed value of the non-auto side (or the non-ignored side in the
> overconstrained case) is whatever normally happens for computed values for
> the property.
> I agree that it should.
> -Boris
Received on Thursday, 12 February 2009 21:02:12 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:38:24 UTC