W3C home > Mailing lists > Public > wai-xtech@w3.org > May 2008

RE: [html4all] HTML5 Alternative Text, and Authoring Tools

From: John Foliot <foliot@wats.ca>
Date: Wed, 14 May 2008 17:23:34 -0700
To: "'Henri Sivonen'" <hsivonen@iki.fi>, "'Christophe Strobbe'" <christophe.strobbe@esat.kuleuven.be>
Cc: <wai-xtech@w3.org>, "'Dan Connolly'" <connolly@w3.org>
Message-ID: <009c01c8b621$ea1ec440$e63142ab@stanford.edu>

Henri Sivonen wrote:

> I think in order to properly analyze what's going on, we need to set
> aside how we wish the world to be and examine how it is.

More images today have some decent alt text than if @alt had remained
optional (as it was in HTML3.2).  Many images have alt="", and many of those
are probably wrong, although some are probably right.  Mandating @alt in
HTML 4.01 has caused more page authors to think about the issue, and coupled
with outreach from web accessibility advocates, has probably improved the
state of affairs, if even marginally.  Finally, insisting on alternative
text for all images is one of the most basic of "web accessibility" concepts
we have to sell our message.

> I didn't say
> alt is bolted on for the sake of disagreeing. I think the realization
> is crucial for understanding why alt continues to be an issue. There
> are accessibility features with which you author once for all
> modalities. These don't even feel like accessibility features, because
> you get accessibility for "free". These features have accessibility
> built in. Then there are features that require dual authoring because
> authoring once doesn't cover some mode of presentation/interaction.
> These features bolt accessibility on.

John's analysis:
Currently, @alt feels "bolt-on" to because it requires additional input from
the content owner to be *most useful*.  However, as others have pointed out,
there is always some additional information attendant to any image on a web
page/document, so the real question might be what if any of that information
is useful enough for users not seeing the actual image.  

However, having @alt mandatory (as it is right now) places it somewhere else
on the continuum , as it is a *requirement* for syntactically correct usage,
not a "bolt-on" per-se (attributes such as height, width, style are
bolt-ons).  HTML WG feels that because of this it is degrading the user
experience (as users will "game the system") and suggest that the best way
forward is to separate the requirement from "conformance" (under special
conditions which cannot be programmatically determined), whilst
accessibility advocates claim that it is shifting the continuum backwards,
and will most likely have negative effects unforeseen, not the least of
which will be more images, not less, appearing on the www without
appropriate contextual information.

> The trouble is getting people to
> do dual authoring. Therefore, other things being equal (and they
> rarely are), we should favor feature designs that you author once for.
> This is what HTML5 is doing with e.g. <progress> and Web Forms 2.0.

Fair enough.  Go back then and come up with an alternative means of ensuring
that when images are present on a page, that appropriate contextual
information is provided - always - else the page be nonconformant.  If not
@alt, then how else?  I for one am open to other suggestions, but you must
*really* understand our concerns as well as the real problem from our
perspective.  Taking away the only actual mechanism we have today that is
*mandated*, and not replacing it with something better (AND *REQUIRED*, NOT
OPTIONAL) is a net loss for accessibility - full stop.  To argue otherwise
is to not understand the issue from the accessibility perspective, no matter
how well meaning your intention may be.  If content authors insist on gaming
the alt attribute for conformance sake, then shame on them, but if the draft
authors make it *easier* for those same authors to skirt their
responsibility, then shame on you.  I know you say that "shaming" will not
drive the discussion further, and I tend to agree, but until such time as
the HTML WG comes back with a real solution to the problem (and not the
current non-solution), there will continue to be disagreement and no forward
movement.

> 
> Alt is one of the things that require dual authoring and images will
> need it for the foreseeable future in order to be accessible. But alt
> is such an issue precisely because it's bolt-on. Vehemently asserting
> otherwise doesn't help in understanding what's actually going on and
> in making language design decisions.

Vehemently asserting that making @alt optional will improve user-experience
doesn't help in understanding that it is not a benefit to accessibility, no
matter how you choose to think otherwise.  If the current implementation is
flawed, repair or rebuild, do not remove.  *THAT* is the real bottom line.
(IMHO)

> 
> (Aside: alt is also bolted on in the sense that it wasn't part of the
> original design of the <img> element.)

Uhm... I (and history) disagree: 
Since HTML 2, the @alt attribute has been part of the specification,
although in 2 and 3.2 it was not mandatory.  See:
  http://www.w3.org/MarkUp/html-spec/L2Pindex.html#IMG (HTML 2)
  http://www.w3.org/TR/REC-html32#img (HTML 3.2)

...and Dan Connolly was the author of record for the W3C html 1 spec... Dan,
correct me if I'm wrong but <img> was not part of that spec was it?

JF
Received on Thursday, 15 May 2008 00:25:02 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 13:15:49 GMT