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

Re: 'initial' | 'inherit' inconsistency

From: Eric A. Meyer <eric@meyerweb.com>
Date: Sun, 31 Oct 2010 12:35:08 -0400
Message-Id: <a06230906c8f34626ddc2@[192.168.1.128]>
To: www-style@w3.org
At 1:48 AM -0700 10/31/10, fantasai wrote:

>>  (2) Which properties do not allow initial and/or inherit?
>>
>>  If the answers are anything but (1) "no, yes" and (2) "none"
>
>I believe the answers are indeed (1) "no, yes" and (2) "none"

At 1:58 AM -0700 10/31/10, fantasai wrote:

>The 'initial' keyword is introduced by the CSS3 Values and Units
>specification. Unless a UA implements that module (which, I remind
>you, is a WD only and not CR), it will not implement 'initial' for
>any properties.

    Maybe I'm just stupid, but don't those points contradict each 
other?  That is to say, "Values and Units" is as you say not CR. 
Therefore how can we say that the number of properties that do not 
allow initial and/or inherit is none?  For that matter, how can we 
even assert the inverse?
    Whereas if all permitted keywords were explicitly listed in 
property definitions, there would be no confusion as to which 
properties do or don't accept this or that keyword.  If a module 
failed to include 'inherit' on one property out of many, questions 
would arise as to why and it would either be explained or corrected. 
If a module chose to add 'initial' or 'default' or 'automagic' or 
whatever to its properties, it could do so with a revision.
    This does mean that in the rare cases where a new "universal" 
keyword is devised, each module would have to be revised to add it, 
but that seems a good thing: it would mean more editors and 
interested parties would have to consider the pro and cons of adding 
the new keyword.

>The 'inherit' keyword was introduced in CSS2.1, and thus any CSS3
>drafts that do not include it are in error.

    Certainly.  But I assert even more fundamentally that any modules 
(draft or otherwise) that do not explicitly define all property 
values will give rise to error.  I think many statements in this 
thread already provide evidence of that.

-- 
Eric A. Meyer (eric@meyerweb.com)     http://meyerweb.com/
Received on Sunday, 31 October 2010 16:35:45 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:33 GMT