Re: Is the px unit equivalent to user units?

Dean Jackson wrote:
> 
> 
> On 04/01/2006, at 5:00 AM, Ian Hickson wrote:
> 
>>
>> On Tue, 3 Jan 2006, Jon Ferraiolo wrote:
>>>
>>> I think nearly all of the long-standing members of the SVG WG would
>>> agree that it would be preferable to drop CSS units from all version of
>>> the SVG language (except for defining the intrinsic size of the graphic
>>> via width/height on the root svg element) rather than require an
>>> abomination such as requiring 'uu' or 'csu' on SVG length values just
>>> because CSS purists have a religious issue about unitless values.
>>
>> I really don't understand the problem. This has nothing to do with
>> religion, it's to do with real problems with syntax (e.g. line-height
>> handles <number> and <length> very differently).
>>
>> We have many times suggested a simple solution:
>>
>>  * Continue to have the "px" unit in text/css be exactly equivalent to
>>    unit-free "user units" in SVG.
>>
>>  * If other units are used, convert them to "px" units (and thus user
>>    units) in exactly the same way that CSS already requires.
>>
>> This is what the specs *already say*. This is scalable. It is completely
>> compatible with all existing specs and content.
> 
> I agree with this solution. (I actually thought we'd already decided to)

I've not seen that said in public, although I certainly could have missed it. An 
official "on behalf of the SVG WG" statement to this effect would be very 
helpful for implementors.

Will you also be saying that the presentation attributes should add a 'px' to 
unitless length values before they are mapped into the CSS style system? I think 
this is important to make sure the value inherited or returned by 
getPresentationAttribute is consistent across implementations. Somewhere around 
the section starting:

   For user agents that support CSS, the presentation attributes must be
   translated to corresponding CSS style rules according to rules described in
   section 6.4.4 of the CSS2 specification

in section 6.4.

Please consider this email to be an official "Issue" I've raised for an errata 
item, where a statement at the earliest possible time on the WG's position would 
be very helpful for implementers.

Thanks Dean.

> Dean
> 
>>
>> Then:
>>
>>  * Remove the requirement that SVG places on CSS parsers to accept
>>    <length>s that are just <number>s.
>>
>> This resolves the one issue that keeps being raised. SVG Tiny 1.2 already
>> does this (in fact it removes all requirements on CSS parsers).
>>
>> --Ian Hickson               U+1047E                
>> )\._.,--....,'``.    fL
>> http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
>> Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
> 
> 
> 

Received on Wednesday, 4 January 2006 18:12:52 UTC