WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: Alt text (was VIKI - text transcodeing)


From: John Foliot - Stanford Online Accessibility Program
Date: Jan 22, 2007 12:20PM

Patrick H. Lauke wrote:
> John Foliot - Stanford Online Accessibility Program wrote:
>> Then why bother with the image at all?
> Because the web is still, for the majority of users, a visual medium.
> Do you offer alternatives for the choice of colour of your background
> or links? No? Then why bother with colour at all? Same reason.
>> To *not* offer the equivalent, is to
>> me, doing them an injustice:
> Same for colour, choice of typeface, layout, etc? They all serve to
> set
> the mood...

Except... If done correctly (via CSS), all of these colour/typeface/layout
choices can be modified by the end user to adjust to the end user's needs
and expectations. So while we may set out to create a mood through visual
design, the end user can changes those choices at whim. However, graphical
elements embedded into a page differ from CSS styles significantly, and
IMHO, should not be grouped into the same category. The expression, "A
picture is worth a thousand words" comes to mind...

>> surely Patrick you are not suggesting that 60%
>> of all images on the web today (used to establish mood) should simply
>> have alt=""?
> No, I'd suggest sticking them in via CSS ;-)

Extraneous flourishes et al. I agree with. A photograph of a teacher in a
classroom - sorry, that is an actual photo, included for a reason (even if
to some it may be trivial).

> As ever in the discussion on alt, this is very much dependent on
> context. And - as I already remember arguing with Joe Clark ages ago
> when he was putting together his canonical reference for when a PDF is
> appropriate - I'm after all those images that are not there to
> illustrate specific things that could be classed as content
> themselves,
> but those wonderfully generic "man with laptop", "businessmen shaking
> hands" cr*p that abounds in corporate design.

Comments aside regarding stock photography, those generic "smiling
businessman" photos, (or the now infamous "two shaking hands") reinforce the
copy. Yes, it may be visual fluff to you, but it is there non-the-less, and
to not acknowledge it is to apply a different standard to those who do not
"see" the image. This comes back to one of the comments I made earlier -
some argue that the ALT text should be nothing but "factual data": alt="two
men handshaking", whereas I have always used the "why it there?" rule, and
have expanded on the alt text to include some context: alt="[Photo - a
handshake: we are committed to ensuring your satisfaction]"

>> Why not: <img src="path" alt="[Photo - Students enjoying the relaxed
>> teaching style of Professor Jones]" />
> If the main copy already has text along the lines of "Professor Jones'
> teaching style is relaxed, and every year students comment on how
> enjoyable his lectures are", then it's pure duplication. All this
> would
> say to me is: look, we've got photographic evidence!

OK, and??? Why would you indicate this to a visual user, and neglect to
inform a non-visual user (or simply someone who has disabled images for a
specific reason)? The fact that you are illustrating "...we've got
photographic evidence" is in itself an important statement don't you think?

Why should we, as enablers, also decide what some users should and should
not get? While the role of "editor" is somewhat inherent in what we do
(witness my alt text above), I don't think we should be deciding what is
appropriate or not for any user; rather I think we should be striving to
ensure that the spirit of the original content author is respected, and
conveyed as completely and truthfully as possible to all users.

> As mentioned above, this also depends on the context of the
> page/image.
> I'm crusading against the purely "visual fluff" imagery. If your image
> above was in a generic page about teaching at an institution (rather
> than a page specifically for Prof Jones), then it probably falls more
> under content.

I understand the point you are advocating, and I think we can agree on the
idea of context; however I disagree on the general thrust of your argument
that we as developers/enablers start to apply filters to the content. If a
"designer" chooses to use some generic image-bank fluff, savvy users of all
stripes will "know" what that generic photo "means", and will apply their
own internal filters to weigh the importance of the image. I do not think
it is our role to be arbitrator in this regard to those users who may not
"see" the image - we should be more "reporters" than "editors" in this case.
The "functional equivalent" argument to me applies both ways - we offer up
both the good and the not-so-good equally.

>> Your main point of ensuring that the supporting text also supports
>> the tone, mood or feel is 100%, but if the copy-writer wants an image
>> there to support that mood, then the alternative text should do so as
>> well.
> It's not the copywriter that wants it there, it's the designer (who
> will, more often than not, stick some bland image from an imagebank in
> there).

And that too sends a message.