W3C home > Mailing lists > Public > public-respimg@w3.org > September 2013

Re: Compressive images test

From: Marcos Caceres <w3c@marcosc.com>
Date: Tue, 17 Sep 2013 15:50:32 +0100
To: Barry Latimer <blat001@gmail.com>
Cc: Andy Davies <dajdavies@gmail.com>, "public-respimg@w3.org" <public-respimg@w3.org>
Message-ID: <61B74F2F13B643A3AAFF42ED5DB22BFE@marcosc.com>
Hi Barry, 

On Tuesday, September 17, 2013 at 3:15 PM, Barry Latimer wrote:

> All,
> 
> While its really interesting to understand how different image formats compress and byte boundaries work, it seems to be a different topic than responsive images and the ideal implementation method, as using optimal or non-optimal compression techniques to serve different images to a different breakpoints/screen sizes/screen formats should result in different image sizes therefore reducing/increasing the overall download size of a page depending upon the chosen image.
I'm sorry, but I have to disagree. Much of the pushback that we got originally from browser vendors was "compress your images better, then!". In particular, the community group was pointed to: 
http://www.netvlies.nl/blog/design-interactie/retina-revolution

if we are able to show that scaling down giant images (regardless of them having low kilobytes) simply won't scale memory-wise - that's a pretty big deal.  
> The focus here needs to be more around which techniques will best work for the wider community on specifying a format/syntax to indicate to the browser which image to download based upon the screen/user/device viewing the web page in question.

Sure, that is just one of our problems - but the CG is interested in _all_ aspects of image delivery. For instance, we are also extremely interested in evaluating how Client-Hints could help (which is a HTTP-based solution, rather than a markup solution). 

In other words, the problem of effective responsive image delivery is not confined to syntax and formats. It involves looking at a range of problems that affect different layers of the platform (from network, to markup, to the preload scanner, to decompression speed, to memory usage, etc.)

Kind regards,
Marcos 

  
-- 
Marcos Caceres
Received on Tuesday, 17 September 2013 14:51:04 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:12:40 UTC