Re: img@relaxed CP [was: CfC: Close ISSUE-206: meta-generator by Amicable Resolution]

Hi Ben,

You appear to have been selective in your quoting from the spec, directly below your "might" quote is the following:

In a conforming document, the absence of the alt attribute indicates that the image is a key part of the content but that a textual replacement for the image was not available when the image was generated.


On 4 Aug 2012, at 15:29, Benjamin Hawkes-Lewis <> wrote:

> On Sat, Aug 4, 2012 at 1:01 PM, Steve Faulkner <> wrote:
>>> The absence of @alt is an unambiguous indication that the image should
>>> have a text alternative. Doesn't necessarily mean it's key to
>>> understanding the content though.
>> While you may disagree, that is what the HTML5 and HTML LS currently
>> define it as.
> I was talking about what developers can effectively communicate with,
> and what UAs can (and do) reasonably assume from, its absence, rather
> than merely what the spec says about it, but in any case it agrees
> with what I was saying AFAICT:
> "If the src attribute is set and the alt attribute is not … The image
> might be a key part of the content, and there is no textual equivalent
> of the image available."
> It says "might be" not "is". This is the correct inference, I feel.
>> When an image is the content of a figure element it is not unabiguous
>> <figure>
>> <img>
>> <figcaption>text</figcaption>
>> </figure>
> Yeah, I take it as read that we are using presence of @alt as a
> shorthand for presence of @alt or other (non-repair) source of text
> alternative.
>>> User agents (as opposed to linters) have to treat images without @alt
>>> with or without the linter flag the same,
>> why? currently firefox for example does not display any visible
>> indication of an image when it has not alt attribute, why is that
>> useful for users who have images disabled?
> I'm not saying it is or isn't, though I do think user agents should be
> free to render as they see fit.
> I doubt a user agent would factor the presence of an attribute such as
> @relaxed into its decision to indicate or not indicate the presence of
> a particular image without a provided text alternative. In particular,
> I doubt user agents which provide configuration for indicating or not
> indicating such images (such as VoiceOver which allows users to ask it
> to announce all images or only images "with descriptions") would
> distinguish <img> and <img relaxed> as distinct categories in their
> configuration UI.
> Maybe an <img relaxed> would be less likely to be a spacer image. But
> I think there are more reliable heuristics for detecting spacer images
> than looking for the absence of @relaxed, for example looking at
> filename, intrinsic image size, color variance, repetition of the
> image, and legacy traits of surrounding code.
> In practice, spacer images are rare in new content where you could
> treat the absence of @relaxed as a signal. For recent content a bigger
> challenge to alerting users to key content is the abuse of CSS
> background images to speed loading of content images like news photos
> and galleries.
> @relaxed is such a weak signal for user agents that I doubt the value
> of pushing it into the accessibility APIs, that's all.
> --
> Benjamin Hawkes-Lewis

Received on Saturday, 4 August 2012 14:58:10 UTC