RE: [css3-images] image-rendering property for contrast-preserving image upscaling

> From: Tab Atkins Jr. []
> Sent: Monday, February 07, 2011 9:23 AM
> To: Sylvain Galineau
> Cc: Lech;
> Subject: Re: [css3-images] image-rendering property for contrast-
> preserving image upscaling
> On Mon, Feb 7, 2011 at 8:33 AM, Sylvain Galineau <>
> wrote:
> >> I don't believe there's a good reason to specify an explicit algorithm.
> >> When you know precisely what algorithm you want, you can implement
> >> that yourself easily in canvas.  When you don't care all that much
> >> (which should be the common case), you can just declare your intent
> >> and let the browser do what it wants to make everything look as good as
> possible.
> >
> > If the answer for your average web author is 'code your image scaling
> > algorithm yourself' then I would most definitely argue there is a good
> > reason for this ! That is not something the kind of JS code browsers
> > should spend cycles running on a routine basis.
> The average web author isn't going to care, imo.  The 'optimize-contrast'
> value gives sufficient intent to cover the one case in which they would
> (scaling up pixel-art images, as opposed to photos).

Well, someone disagrees and seems to know others do as well. I don't 
see any reason to dismiss their claim at face value so it's worth
investigating a little imo.

> By the time you're caring about the precise scaling algorithm, you're
> (a) advanced enough that you know how to do it yourself, and (b) very
> likely coding an app already, so running the scaling yourself isn't a
> burden.
> > (The implied claim that this can 'easily' be done also begs the
> question:
> > for whom ? And even then, so what ? It's easy to make progress bars
> > and sliders but those are built in HTML5)
> >
> > So the question imo is: why would this be something authors would want
> > to control on a routine basis ?
> Indeed.  I don't think it is, thus my design.  ^_^

It's clear you don't but until you have innate real-time knowledge of author 
needs worldwide asking more questions from those authors who believe they need
this capability might yield more useful input than telling them 'you don't need 
this but if you do, just write the code'. 

Received on Monday, 7 February 2011 18:09:48 UTC