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

Re: Checkpoint 3.4 again

From: Sean B. Palmer <sean@mysterylights.com>
Date: Sat, 28 Jul 2001 23:58:40 +0100
Message-ID: <036a01c117b9$2c6fd640$39206cd5@z5n9x1>
To: "Joe Clark" <joeclark@joeclark.org>, <w3c-wai-gl@w3.org>, "Anne Pemberton" <apembert@erols.com>
[Long-ish rant folows, including a new proposal for CP3.2.]

Hi Anne,

> No matter the arguments presented, text is an element on a
> page, no more, and no less. It needs an equivalent. [...]

This bit of text baffles me (perhaps it needs an equivalent). I
presume the first sub clause of the sentence is a typo, but I'm still
baffled as to what you mean by "text is an element on a page". By
page, I'm presuming you mean document, and by document I'm presuming
you mean some type of SGML format, (e.g. HTML). Usually, when people
refer to the term "element" with respect to Web documents, they mean
the '<img src="a" alt="a"/>' bits. So perhaps it is just my stubborn
vocabulary that is getting in the way.

The case that Joe presents is compelling. There are some things that
are expressible in text that cannot be conveyed by images, and vice
versa. You can often supplement these media types, but you can't
really provide equivalents. Even if you have a simple banner with some
text in it, it is very difficult to get across what that banner looks
like to people, because you will have to interpret its color, font
style etc.

The Pemberton Principle is something that people seem to have missed
for some time, and something that I for one have been very guilty in
not following. But when you take "illustrate where this may benefit
your users" to the extreme, I feel that this damages the case that you
have built up for it. Perhaps after struggling for so long to get
people to realise the "illustrations are acceptable for all, useful
for some, and essential for others" you're trying to subject us to the
absurdities that we've subjected in the opposite direction for so long
:-)

I support your (and Joe's) comments that WCAG 2.0 is ambiguous in what
it says about providing illustrations. It does not phrase it well at
all, in my humble opinion. It is clear that on some occasions,
providing images is essential because it puts across concepts that are
almost impossible to process. A good example is when explaining RDF
Schemata principles to people using XML RDF code; it's very difficult
to read. Node and arcs diagrams on the other hand are simpler for most
people to grok. People who can't see the nodes and arcs diagrams? Once
the problem of how to gracefully transform SVG is sorted out, it
should be easy to tab through or otherwise explore diagrams in ways
which would simply not be possible for bitmapped images.

So to WCAG 2.0: as with all guidelines/techniques documents, the
layout is bound to be inherently flawed due to the fact that it
interprets and then displays the concepts that we are trying to
express: a hierarchy instead of a Web. We also know that certain forms
of hierarchy are more effective than others, although the exact
details of the relationships are subjective. Let's not quibble over
this waste of time, and instead concentrate on what we really want to
get people to do.

We want people to illustrate text, if they have the tools to do so,
where this would benefit the user. We must realise that sometimes,
providing illustrations is difficult, in much the same way that we
must realise that writing clearly is not easy for some people. I don't
consider myself to be a clear writer, and especially when compared to
the likes of William or Wendy, my writing style could be considered
terrible. We can't force people to compensate for things that are out
of control. To do so would be an incredible absurdity.

What do we hope to achieve by the document stating these facts?
Really? I think that we want people to mull over the fact they, hey,
maybe they aren't illustrating their documents enough. We can't force
them to illustrate. It's even difficult to evaluate the quality of
illustrations. All we can do is give people the idea, and hope that it
has an effect (discernible, or non-discernible) in the way people do
things.

It didn't really come home to me until William was sitting next to me,
and I presented my homepage to him with the comment "note that it has
no images". He simply replied, "oh, you have to illustrate it though"
or somesuch. Bang! Like the koan finally being resolved that it was,
the mystery of the Pemberton Principle came through. I believe that Al
did the same thing to William. I'm not sure how Al came across it.

We want to somehow reproduce this in WCAG 2.0. Trouble is, we can't
all have William sitting next to us when we're designing sites :-)

We don't want people to say "that's absurd" (illustrate everything),
and we don't want them to say "duh" (illustrate something). We want
them to realise the importance of illustrations. Here, then (after all
of this waffle) is my proposal:-

Checkpoint: 3.4 Utilise content in a wide range of modalities where
possible to assist the users of your content.
Explanation: Where tools permit and cases warrant, provide, for
example, illustrations for text that may enable people to pick up on
concepts that would previously be difficult to comprehend.
Rationale: Pointing out concepts, or augmenting current ones using
illustrations or other media is acceptable to all users, may be
helpful to many, and under some circumstances may be imperative for
others.

(more specific examples can follow).

I think that gets across in a sensible and realistic manner the
concept that we are trying to impart to people with this checkpoint.
It's short, and to the point.

And please, please, please, get rid rid of the non-normative example:
it's absolutely rediculous. The phrase "a text equivalent equals its
non-text equivalent" is theoretically correct (all eqiuvalents are
equal!) but practically, it's mind boggling. Perhaps people have a
much looser definition of "equivalent" than I do. When something is
equivalent to another, in the mathematical sense, they can be
substituted without changing anything that the formula is expressing.
If text were equivalent to images in any way, we're be able to use
them totally interchangable without any side effects. This is clearly
not the case (else we wouldn't be having this discussion, and the
checkpoint would have no meaning!), so although the formula "a text
equivalent equals its non-text equivalent" is logically sound, there
can be no instances to justify it.

Cheers,

--
Kindest Regards,
Sean B. Palmer
@prefix : <http://webns.net/roughterms/> .
:Sean :hasHomepage <http://purl.org/net/sbp/> .
Received on Saturday, 28 July 2001 19:04:06 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:47:11 GMT