W3C home > Mailing lists > Public > w3c-wai-eo@w3.org > January to March 2009

Re: null alt text

From: William Loughborough <wloughborough@gmail.com>
Date: Thu, 5 Feb 2009 07:03:15 -0800
Message-ID: <1e3451610902050703u79ee6eeeie2207f92ab2b72d3@mail.gmail.com>
To: Liam McGee <liam.mcgee@communis.co.uk>
Cc: EOWG <w3c-wai-eo@w3.org>, "Gregory J. Rosmaita" <oedipus@hicom.net>
So the classic argument becomes "But what if the choice is between terrible
poetry and silence?" That choice should be pre-settable within the screen
reader, not an escape hatch for authors who suck at providing an alternative
to any unwanted voicings. It's all in the hands of the authoring tool and
the design of the assistive technology.

In sum, we are not providing an "equal" experience for the screen reader
user in deference to the "until the technology gets smarter" offput of
responsibility. If you are going to provide "accessible" content (and not
doing so makes you a scofflaw!) you should leave no element unaccounted for.

We have skated around this matter for over a decade and the usual arguments
against full inclusion continue to plague us. It's back to the old "why
would a blind guy want to know where/why I put that <hr>" or how the page


On Thu, Feb 5, 2009 at 3:38 AM, Liam McGee <liam.mcgee@communis.co.uk>wrote:

> William Loughborough wrote:
>> Me and Gregory have been about the only voices in WAI forums for using alt
>> Now a survey at http://webaim.org/projects/screenreadersurvey/#imagesseems to validate our view because this attitude is shared widely by PWD but
>> NOT by non-disabled observers. Also, it seems the more skilled at using
>> screen readers the user is, the more she prefers information to any concern
>> with repetitive stuff which they apparently can filter out with little
>> trouble.
>> My contention remains that if one wants to avoid over-verbiage it should
>> be a function of the screen reader NOT the content author.
>> Love.
> Hi William,
> very interesting. Yes, in my experience user testing with partially sighted
> users, there is certainly a wish for information about images that can be
> seen on the page but not made out visually by the user.
> However, in my purist's soul there is a difference between a description of
> an image and an alternative to that image. I take the view that the alt text
> should be functionally or aesthetically equivalent to the image it supports,
> but that it should not usually be a description of the image. We have the
> title and longdesc attributes for that.
> So, to reiterate: alt is an equivalent in a different modality. It is not
> metadata. It is co-equal.
> So when is a null alt appropriate? Well, certainly in the bad old days of
> spacer gifs I wasn't a fan of the alt text value of "spacer". What about
> pretties though? Is it possible to translate from a visual aesthetic to a
> verbal one? The visual aesthetic evokes certain emotional responses... so
> should we seek to evoke similar emotional responses with the alt text, so
> that it is a true alternative, just one in a different modality? Perhaps alt
> text should be poetry, with mundane description reserved for title and
> longdesc. But what if the choice is between terrible poetry and silence? I
> would argue that in this case, silence is golden, and if you want the
> description of the visual effect of the image, get the title of the element,
> not the alt.
> Just my 2p.
> Regards to you all, see you Friday, in an aural modality.
> L.
> --
> www.communis.co.uk
Received on Thursday, 5 February 2009 15:03:52 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:55:57 UTC