RE: [css3-images] aliases for 'cover' and 'contain'

I know you were asking Daniel, but I'll chime in anyway. ;)

"unless we have a serialization strategy that preserves the right alias"

I completely agree that contain, closest-side, cover, and farthest-corner should be preserved independently from an OM and serialization perspective, even though they have the same rendering result.

Otherwise, I think the "let's remove them" plan makes more sense.

-Brian

-----Original Message-----
From: Florian Rivoal [mailto:florianr@opera.com] 
Sent: Monday, November 07, 2011 11:26 AM
To: www-style@w3.org
Cc: daniel.glazman@disruptive-innovations.com
Subject: Re: [css3-images] aliases for 'cover' and 'contain'

On Mon, 07 Nov 2011 18:15:24 +0100, Tab Atkins Jr. <jackalmage@gmail.com>  
wrote:

> On Sun, Nov 6, 2011 at 9:58 AM, Brad Kemper <brad.kemper@gmail.com>  
> wrote:
>> I vote that we just keep the aliases. I anticipate that most people who  
>> use the keywords will alternate between 'cover' and 'contain'. But  
>> those who do want to use farthest/nearest-side/corner would probably  
>> prefer them to be all consistently named. The computed value would just  
>> be 'farthest/nearest-side/corner', which could be a problem for  
>> round-tripping, but still seems like a reasonable trade-off.
>
> Given the way the syntax discussion turned out, I agree.  Fantasai's
> example usage "radial-gradient(from 25px 25px to cover, blue,
> transparent)" works just fine if we keep the current definition of
> "cover", since the "as" clause defaults to center when omitted.

This logic is sound, but at the same time, I am a bit worried about  
serialization converting one alias to the other, and getting the same kind  
of complaint we get about red turning into rgb(255,0,0).

I suppose the ones most inconvenienced by stuff like that would be editors  
and such things, as they try to preserve not only the semantics, but also  
the actual syntax written by the author.

Daniel, am I correctly anticipating you disagreeing on having such aliases  
unless we have a serialization strategy that preserves the right alias, or  
do you actually don't mind?

  - Florian

Received on Monday, 7 November 2011 19:31:34 UTC