[whatwg] <imgset> responsive imgs proposition (Re: The src-N proposal)

Bruno Racineux bruno at hexanet.net
Sat Nov 16 20:34:49 PST 2013

On 11/16/13 3:24 PM, "matmarquis.com" <mat at matmarquis.com> wrote:
>I started and have been running the Responsive Images Community Group for
>nearly three years now.I helped build BostonGlobe.com and the original
>³prototype² script for responsive images, I¹m a maintainer on the
>Picturefill project, and I wrote large parts of the `picture` proposal. I
>have a pretty fair idea of the current ³best practices² for responsive

You are taking it personally on a misread, and my admittedly exaggerated
verbose to criticize your counter-argument. By 'current practices', I was
solely referring to the average length of image paths, which you gave me
no factual argument other than "I¹d say the likelihood" is ³very slim.²
with non practical examples.

I was partly aware of your background, but that's besides the point of
this argument.

Please just don't qualify my example as 'exaggeration' with a fictitious
'more common representation of usage' in term of average char length.
The average src path of an img on the web is between 50 to 70 chars.

That's the sole point of my argument here, as per my examples.

And I don't know why you seem to talk about 'hotlinked' images as a
special category... I just hope that your argument is not that
Hotlinked/Cloud/CDN images don't apply to src-N or srcset, which is what I
first read.

On 11/16/13 3:15 PM, "Yoav Weiss" <yoav at yoav.ws> wrote:

>Claims of research and facts are usually more credible when linked to an
>actual research and/or facts.

Yoav, I didn't feel the need to point to non-existing studies to prove
a fairly obvious point, other that citing something a dozen examples
mentioning with wordpress + drupal which roughly cover 50% of cms usage
stats, or over 30% of the web.

But if you must have those particular facts, here they are:

> On 11/16/13 3:24 PM, "matmarquis.com" <mat at matmarquis.com> wrote:

>The examples above are entirely examples of sites that use templates for
>their output, at which point things have effectively left
>³hand-authoring² territory. In authoring a page, we¹re building and
>handing off those templates for integration with a CMS, at which point
>those sources are output on the backend. I can¹t think of many instances
>where I¹ll be forced to hand-edit the *output* of a Drupal site.

Neither do I. I am talking about the resulting size of an HTML page, and
readability of src-N code in the broad context of debugging.
Not code editing per say. I don't recall saying that the code had to be
editing friendly.

>A more accurate example in terms of actual authoring with any of things
>you¹ve cited above‹specific variable syntax notwithstanding‹would be
>something like:
>src-1="100% (30em) 50% (50em) calc(33% - 100px);
>          {img1} 100,
>          {img2} 200,
>          {img3} 400,
>          {img4} 800,
>          {img5} 1600,
>          {img6} 3200"

I am not concerned about the authoring of it. Not sure why you are
referring to authoring with that example? I am actually not sure you are
getting my point about 'bloat' and whether your couter-argument actually
talk about the same thing...

>But for the sake of discussion, I¹m fine with inflating any syntax
>pattern up for discussion equally with hotlinked Flickr URLs, since
>there¹s really no point in arguing either way. ³Six of one, sixty
>characters of another,² as they say.

More information about the whatwg mailing list