W3C home > Mailing lists > Public > wai-xtech@w3.org > August 2007

Empty vs no alt attribute (was Re: Baby Steps or Backwards Steps?)

From: Sander Tekelenburg <st@isoc.nl>
Date: Thu, 16 Aug 2007 17:41:41 +0200
Message-Id: <p06240603c2ea13af33ff@[]>
To: HTMLWG <public-html@w3.org>, wai-xtech@w3.org

At 19:43 +1000 UTC, on 2007-08-16, Jason White wrote:

> On Wed, Aug 15, 2007 at 10:52:44PM -0700, Maciej Stachowiak wrote:


>> 1) Photo sharing sites like flickr.com. It would be wildly impractical for
>> such a site to prompt the user for alt text for every image [...]
> They could supply alt="unknown" and allow the uploader to submit a properly
> descriptive ALT text,

I cannot imagine how alt="unknown" could be useful to anyone. Can you
explain? What exactly can a user derive from that that is useful? What
exactly does it tell that a non-existent alt attribute does not? (Maybe I
simply misunderstand what you mean?)

Note that ATAG 1.0 explicitly states that authoring tools must not insert
canned @alt values: "Do not automatically generate equivalent alternatives.
Do not reuse previously authored alternatives without author confirmation,
except when the function is known with certainty." [See


>> 2) Mail clients that generate HTML. A user may be inserting an image or
>> multiple images through drag-and-drop or copy/paste. Again it would be
>> impractical and annoying to prompt the user.

FWIW, the mail context seems a different one to me. On the Web, authors
cannot know the user's browsing environment. With email, you often do know
something about the recipient (not on mailing lists). So ideally an email
client would differentiate: don't bug the author for equivalents when he
knows the recipient doesn't need or even want them.

> It would be neither impractical nor improper to prompt the user in such a
> case.

Reality is that many authors will simply get too annoyed by being prompted at
'everything' they do (they already say "OK" to any virus that asks if it may
infect the system ;)). A good UI guides users to 'automagically' do 'the
right thing' (I'm not saying that's easy). ATAG 1.0 is aware of this.
Although it says "Prompt the author to provide equivalent alternative
information." [see <http://www.w3.org/TR/ATAG10/#check-provide-missing-alt>],
it also says "Subtle, rather than extensive, prompting is more likely to be
effective in encouraging the author to verify accessibility where it cannot
be done automatically." [See <http://www.w3.org/TR/ATAG10/#def-check-for>]

Some users will want to be prompted at every corner. Others will just get
annoyed, blindly hit "OK" buttons, or move to a less annoying tool. Thus I
agree with ATAG 1.0: "[...] the tool may prompt the author to verify that
equivalent alternatives for multimedia are appropriate. This is the minimal
standard to be satisfied." [See <http://www.w3.org/TR/ATAG10/#def-check-for>]
In other words: the minimum is to ask the user before saving the document, if
appropriate equivalents have been provided. (And such a dialog should include
a link to relevant (built-in) documentation, and to (built-in) tools that
help assess the quality of those equivalents.)

I'd go even one step further: the minimum would be to prompt before
*publishing* a document. Not per se upon saving it. (Some authoring tools
allow saving drafts; or allow some users to edit content, but require a
senior editor's approvement before the content actually gets published.)

I do think that such an authoring tool should have the built-in option to
immediately prompt for every case. That's good for those authors who prefer
to work that way. But it's not going to be effective to force that upon

> Furthermore, if the image already has a textual description or
> equivalent in metadata, it is easier to have it supplied as a value of @alt
> the authoring side than to require the user agent to download the image and
> extract it.

Are there image formats that can contain equivalents? Are those formats used?
Do authors enter equivalents in such image files? If so, I agree that the
authoring tool should grab that equivalent. However, it should still require
user confirmation, because the image may well have come from another site and
thus may copntain alt text that imade sense *there*, but not automatically
anywhere else. So by "grabbing" I would mean "pre-select that", but still ask
for confirmation.


Sander Tekelenburg
The Web Repair Initiative: <http://webrepair.org/>
Received on Thursday, 16 August 2007 16:16:32 UTC

This archive was generated by hypermail 2.4.0 : Friday, 20 January 2023 19:58:27 UTC