Re: WebP, anyone using it?

I've had 6 or 7 requests for WebP support in ImageResizing.Net. Right now it seems to be a chicken-and-egg situation. With no CDN-friendly browser sniffing solution, and no prefetch/noscript compatible client-side solution, I don't anticipating image formats ever being able to progress unless we solve this problem with a type attribute.

Browser-side implementation difficulty is extremely low in this case; it's string comparison against a set of supported strings.

The primary argument against putting this into the spec as an optional attribute is how to handle multiple density images of different types. I say - don't. There's no good reason to need a different image format for a higher res image, and unless the type attribute is specified, multiple formats can be supported.

The other alternative is going with the 1-1 element-url approach, which does sacrifice a bit of compressible verbosity for the sake of simplicity.

Any compromise that leaves the door open to new image formats is worth it, IMHO. It may even be more important than the responsive side of our work here.

Mime-type sniffing and wasted requests is not a solution acceptable to any dev I know of.

-Nathanael

Sent from my iPad

On Oct 15, 2012, at 10:00 AM, Marcos Caceres <w3c@marcosc.com> wrote:

> 
> On Monday, October 15, 2012 at 4:56 PM, François REMY wrote:  
>> By the way, I don't really grasp why the RespIMG spec should talk about
>> image formats. This is very controversial and could create a lot of high N/S
>> debates unrelated to the core issue (which is: responsive image).
> 
> The point is not to discuss particular formats. It is that, if as a developer, I have a smaller/faster/better image in some format, I should have some way of including it in the list of sources without needing to do browser detection.   
> 
> Having said that, this falls in the "nice to have" category.  
> 
> --  
> Marcos Caceres
> 
> 
> 
>

Received on Monday, 15 October 2012 16:24:13 UTC