Re: [css-style-attr] SVG WG comments on CSS Styling Attributes Level 1

I think we should also remember that it is the CSS users that are hurting here (or "humans" as Håkon called them). Not being able to roundtrip is extremely confusing to them, and frustrating to work around. We should be trying to make their life easier.

Dean


On 02/09/2010, at 6:58 AM, David Singer wrote:

> I really still don't see the problem here.  It's reasonable to want to round-trip numbers through javascript.  It's reasonable to want uniform number handling in W3C specs, especially between SVG and CSS.  E-notation parsing is, for the most part, considered a solved problem in reasonable code size, I believe :-).  Unless it introduces problems stronger than 'don't like' / 'don't need' (e.g. the presence of the 'e' seriously complicates other parts of the grammar), we should introduce it and move on. And spent the braincycles of the group on Really Interesting Problems (TM). Please?
> 
> On Sep 1, 2010, at 13:30 , Håkon Wium Lie wrote:
> 
>> Also sprach Boris Zbarsky:
>> 
>>>> (I also think there should be a way to serialize JS without suddenly
>>>> switching notation.)
>>> 
>>> In theory, there is.  It will, generally speaking, produce decimal 
>>> numbers with up to 300 digits or so, right?
>> 
>> That sounds exhaustive. There is no way to set the precision? Like "%.2f"?
>> 
>>> Note that JS will only switch to e-notation for numbers whose absolute 
>>> value is more than 10^20 or less than 10^{-6}. [1]  While using decimal 
>>> notation for the latter can sort of be ok if they're close to 10^{-6}, 
>>> for the 20-digit or more ones it gets pretty silly very fast.
>> 
>> Right. I can't think of a meaningful way of using numbers as small as
>> 10^{-6}. At that stage, most properties would be happy with
>> zero, I believe.
>> 
>> -h&kon
>>             Håkon Wium Lie                          CTO °þe®ª
>> howcome@opera.com                  http://people.opera.com/howcome
>> 
> 
> David Singer
> Multimedia and Software Standards, Apple Inc.
> 
> 

Received on Thursday, 2 September 2010 20:18:52 UTC