Re: More on 3.4

On Wed, 1 Aug 2001, Matt May wrote:

  MM  It's a technical limitation of GIFs and JPEGs that metadata can't be
  stored within, which is why alt text is tied into the HTML anchor.

This is not true. For a set of tools that actually does include metadata
within JPEG images see the W3C Note Describing and retrieving photos using
RDF and HTTP (28 September 2000, by Yves Lafon, and Bert Bos) at
http://www.w3.org/TR/photo-rdf/

The tool could be readily adapted to work with GIF, PNG, SVG, WebCGM images,
all of which also allow inclusion of arbitrary data.

future and philosophy stuff:
Because SVG images can easily include other image types, and has good
metadata support, they are a particularly promising part of the puzzle of
linking descriptions, ideas or concepts (encoded in the Semantic Web) and
primarily textual hypertext with images or re-useable image components.

MM
  It's a
  technical limitation of Flash that users can't control the speed of the
  presentation, so WCAG ensures that Flash movies implement control.

CMN In otherwords, it isn't the technology, it is the author producing
particular kinds of content.
MM
  Technology is the means to overcome the user's limitation, and WCAG (and
  ATAG and UAAG) uses technology throughout.

CMN Technology is one method. Another common method is the way content is
expressed (for example producing text equivalents). In general there is a mix
of the two.
[snip]

old CMN We aren't ordering people to do stuff, we are telling them what they can
  > do in order to make sure that their message gets to more of the people. And
  > the more accurately we can explain that, the easier it is for authors to
  > understand what we mean and to put it into practise effectively.

  MM Here's more of that philosophy stuff:
  First, the reason I'm bothered by this is that it's going to be tied to a
  compliance scheme, and when they are adopted as goals in organizations or
  governments, they become de facto rules, and we _are_ ordering these
  changes.

CMN I agree that there are organisations who adopt WCAG wholesale as a
requirement, and then find out it they can't get there quickly. There are
others who adopt it and find they can get there quickly. And there are others
who read it first, and adopt different parts of it at different stages,
because providing complete accessiblity to everyone isn't something they can
do immediately and they decide to tackle it group by group.

It isn't our role to produce stuff group by group, it is our role to describe
how to remove barriers for all people with disabilities, as far as that is
possible. One approach is that outlined by Kynn, of going back to looser
structures for conformance. There are others..

MM
  And yet, I remain unconvinced that rules can or should be made of 3.3 and
  3.4 (though I should underscore that they should be there nonetheless).

CMN
I am not sure that we can make complete rules. I do believe that we can go a
lot further towards those than just the checkpoints we have at the moment
(some more techniques would be good for a start!)
MM
  ... They
  are both pretty firmly in the generic usability realm,
CMN for many people, but for some people they are straight-out issues of
coping with a disability
MM
  ...and nearly all the
  science there is experiential. That's why we test the stuff!
CMN
Agreed. But while nearly all is experiential, some is not. And where it is,
it is still important to test it, and not just decide "a priori" that we
can't do it.
MM
  ..Changes in
  content can often lose as many people as it helps, especially when
  implemented as some golden ratio, and that's why I see this as the wrong
  approach. That is, content that is more accessible for some people with a
  certain disability may be less accessible for others with the same
  disability.

CMN Agreed. I am very strongly with you on the idea that there is no "golden
ratio" or quantitative level assessment that we can usefully set as a firm
rule. So I am arguing that in some cases content should be supplemented - for
example, recorded speech should be supplemented with a textual record of what
is said (and, I would argue, signed captions in appropriate languages,
especially for complex content), graphics should be supplemented with text
descriptions, and text, should be supplemented with pictures, sounds, and so
on, to ensure accessibility of the content to as many people as we can reach.
MM
  If any of the usability divas could create general rules for content, I'd
  guarantee you that the books in the field would be a whole lot shorter (and
  better written!).
CMN
Many of these things do come good in general rules, that can be well-stated.
There are always edge cases (good rules make this apparent) and the first
time people come across ideas it is helpful to have more explanation. And
that is why books on the topic are longer than the quicktips card, and why
saying "provide text equivalents" isn't enough.
MM
  ..But guidelines for content itself are necessarily
  situational. Making rules that apply for all content is an idea as quixotic
  as making a law setting pi to 3.2 -- though that, too, has been
  attempted[1].
CMN
  Agreed. But making rules that say it works to use 3.2 as an approximation
for pi in the following circumstances does work. As I said above, good rules
account for edge cases or limitations of applicability.

  MM Success criteria for a checkpoint like 3.4 belong in a style guide,
  which could be hundreds of pages long.

CMN This is just an assertion about your opinion. In my opinion, I think we
can write success criteria in about the same size that we can do so for
textual equivalent material. Anyway, I think we need to have the information
before we make that decision, and as far as I can tell our charter binds us
to look for that information or attempt to produce it.
[snip]
  MM But all of this is replacement content, meant to fill a hole left by an
  inaccessible portion of the document. It doesn't change the content or flow
  of a document. Images and multimedia augment the content on the page, and as
  such, they do change its format. They are invasive in a way that 1.0 rules
  are not.
CMN Huh??
  old MM
  >   We do assume that the author has the skills and tools to make a site
  >   accessible.
  >
  old CMN For example the ability to write, to interpret graphical content from
  the
  > world around them, to provide text equivalents for sounds of all kinds...

  MM But not to communicate visually. This has been taken for granted
  consistently in this thread as a skill people have, and my experience with
  trained professionals tells me it is not.

CMN No, I am not taking it for granted. I do not assume taht everyone can do
it becuase I know there are plenty of people who can't. The same is true of
writing - there is plenty in the archives of this list that I find almost
impenetrable.

I am not starting from an assumption that anybody can create accessible
content on their own. I am starting from an assumption that there are things
that need to be done in order to make content acccessible, and that we ought
to be working out what those things are, how they can be done, how we can
specify that so people understand, and what tools we can provide to help
people do it.

  old MM
  >   Every rule in 1.0 can be implemented for HTML and CSS using the
  >   same tools and skills the author used to create the site.
  >
  old CMN This is not true. Unfortunately most Authoring Tools make it
  difficult to
  > produce accessible content, although they make it easy for some people to
  > produce content. That's one of the reasons why the work of the Authoring
  Tool
  > Accessiblity Guidelines Group is so important. It is a major goal of the
  Web
  > that anyone author their own content, which means people need tools that
  > support them to do that, and to do it accessibly. (On the positive side,
  > tools are getting better).

  MM I was referring to "skills and tools" as a set not including "Authoring
  Tools". What I meant was that anyone who has access to HTML and CSS source,
  a text editor, and knowledge of those languages can make the changes. The
  difference with 3.4 is that where the 1.0 requirements can be done using no
  more than semi-skilled labor, low technology and no content-development
  skills, and can usually be done after the actual content, suddenly 3.4 needs
  not only graphic design skills and tools, but also usability training (if
  not testing) _and_ content skills, all in one package, and much earlier in
  the design phase of sites than a large percentage of sites are prepared for.
  (And yes, I'm expecting someone to say that's just tough for the sites, but
  if you can't retrofit for accessibility in the late stages of design, you
  can't make much of the content out there today "accessible", nor can you do
  it with repurposed content. The accessible world gets smaller and
  smaller...)

CMN Many current Web authors do not have the ability to work on source code.
That is not a job that can be done by semi-skilled labour, it requires
particular training that is not the same as the training that people who are
producing content actually have.

I can teach my nephews to film themselves playing, plug the camera into my
computer, and put the film online. I can do the same for a 20-year-old I have
met who works as a car cleaner and does simple mechanical tasks. I can teach
them to link this page to another one, so people can find it. But none of
these folks are remotely capable of providing a text equivalent. Masafumi
Nakane takes a lot of photos, and can put them online without any difficulty.
But because he is blind, he cannot provide a text equivalent for them. He
relies on someone else to do so. Examples are infinite, unlike time.

cheers

Charles

  old MM
  >   And authors should
  >   be able to interpret the images down to alt text using their knowledge
  of
  >   both the content and context of the image (that is, we _already_ are
  >   depending on the author's knowledge of his or her content to make things
  >   accessible). What they can't do, reliably, is learn visual communication
  at
  >   the drop of a hat.
  >
  old CMN
  > But then, teaching literacy at the drop of a hat has also proved somewhat
  > challenging...

  MM Textual and visual literacy both. A great many concepts, from
  poorly-illustrated simple ones to brilliantly-illustrated complex ones, will
  remain incomprehensible. The important element here is quality, not
  quantity.

  [1] The Indiana State House attempted to recognize this "mathematical
  truth" -- royalty-free -- in 1897:
  http://www.urbanlegends.com/legal/pi_indiana.html

(In decimal numbering systems it isn't true at any point - it only works for
people who think in roman numbering or similar systems).

Received on Wednesday, 1 August 2001 17:35:45 UTC