W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > April to June 2001

RE: What is illustrated

From: Bailey, Bruce <Bruce.Bailey@ed.gov>
Date: Mon, 7 May 2001 08:29:07 -0400
Message-ID: <5DCA49BDD2B0D41186CE00508B6BEBD0022DAF07@wdcrobexc01.ed.gov>
To: "'A.Flavell@physics.gla.ac.uk'" <A.Flavell@physics.gla.ac.uk>
Cc: WAI Guidelines List <w3c-wai-gl@w3.org>
Dear Alan,
I very much appreciate your perspective on appropriate ALT tag content.  I,
for one, would find additional concrete examples of your philosophy helpful.
Would you please "close the loop" for me by suggesting a better code for
this example:

>  [Picture of George Washington] George Washington in the French and
>  Indian War

Let us assume that the IMG used was one of the more common paintings and
that the particulars are incidental.  It is not clear to me if the sentence
fragment was meant to be a caption (to the portrait) or the start of a
paragraph.  Does this matter?  In either case, ALT="" strikes me as
inappropriate, but I cannot think of an alternative which (1) is
appropriate, but (2) does not interfere with the text-only reading flow.

Thanks, Bruce

> ----------
> From: 	Alan J. Flavell
> Reply To: 	A.Flavell@physics.gla.ac.uk
> Sent: 	Friday, May 4, 2001 6:49 AM
> To: 	Adam Victor Reed
> Cc: 	WAI Guidelines List
> Subject: 	Re: What is illustrated
> 
> On Thu, 3 May 2001, Adam Victor Reed wrote:
> 
> > When reading the text-only presentation, I still want to know
> > what the author considered worth illustrating -
> 
> Indeed; and Anne has already told that to us, when she says (if I
> understood her intentions right) that the illustration was intended as
> a visual reinforcement of the identification that is already present
> in the text.  I would submit that the visual reader is expected to
> give the picture the merest glance, and then move on.  The text
> equivalent should serve the equivalent function, in my submission.
> 
> Let's take another example.  What do I have on the desk in front of me
> right now?  Well, there's a small red portrait of Her Majesty the
> Queen.  Is this a useful answer?  Of course not: the real answer to
> the question is "a stamped envelope".  It is technically accurate to
> say that there is a small red portrait of HM the Queen (namely, the
> postage stamp), but this detail is neither useful nor informative in
> the context.  And so, taking time out to describe this kind of
> (factually accurate but not directly relevant) visual detail would be
> to impair the real message under a mass of detail.
> 
> > It is worth my while to take in a couple
> > more words, as in "[Picture of George Washington]", to tell me
> > what the illustration was.
> 
> The textual content then flows as
> 
>   [Picture of George Washington] George Washington in the French and
>   Indian War
> 
> I find this incongruous, and would prefer to avoid it.  As I've tried
> to stress all along, I'm attempting (wherever possible) to offer the
> text-mode reader a full-valued text document which carries the
> author's intentions - and not a mere second-best description of a
> visual experience.
> 
> I really would have preferred, as I hope I made clear, for any overt
> text caption to a picture to be formulated as a sentence capable of
> standing on its own feet - rather than apparently dangling in need of
> the picture to support it.  Unfortunately, we have no really effective
> way of providing a textual caption that gets suppressed when the image
> itself is not viewed, and so we are stuck with this kind of technical
> detail.
> 
> As I mentioned before, HTML gives you additionally the TITLE and
> LONGDESC (formerly the D-link) for supplementary content where this is
> felt to be relevant.  And some, at least, text-mode browsers give the
> reader the option to enforce presentation markers on images, even
> where the author has supplied empty ALT texts, so that a reader who
> insists on knowing where there are images, even when the author has
> decided they would be a pointless distraction, can have their way.
> 
> While it's true that we are discussing compromises, and borderlines of
> responsibility between the author to set down their message, and
> readers to have that message rendered in a way that is congenial to
> them, I would still like to emphasise the benefits of producing a
> document which in text mode represents a fully-valued text mode
> document in its own right, rather than seeming to be a second-best
> description of something which was meant to be inherently visual.  
> Sometimes, just sometimes, the latter is inevitable: how could we
> discuss a beautiful painting without reference to its visual
> properties?  
> 
> But, often this is not so, and (if I may put this in devil's advocate
> terms) the pictures are often only there for the benefit of those
> whose lack of imagination prevents them from getting the desired
> result from a textual description.  Just think in terms of re-writing
> a TV show for sound radio: would you pester the listeners with
> descriptions of all kinds of visual detail of the TV show, or would
> you just get on with the main business of the show's content?
> 
> 
> [1]It surely goes without saying that images which are standing alone
> as web links absolutely require a working ALT text.  But please, not
> examples like ALT="Red bullet".
> 
> (This whole presentation assumes, of course, that browers are
> available, to those who need them, which support the design intentions
> of ALT, TITLE and LONGDESC.  Compromises may be inevitable while
> waiting for those, but I'd rather not lose sight of the target while
> doing that.)
> 
> best regards
Received on Monday, 7 May 2001 08:29:52 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:47:10 GMT