RE: REF 1.1a - Add definition to 1.1 for ability to be expressed in words

At 10:30 AM 2003-07-11, John M Slatin wrote:
>Gregg, the whole notion of "equivalence" is deeply problematic from any 
>kind of theoretical/philosophical standpoint (to say nothing of the 
>pragmatic issues!).

Yes, "there must be a text alternative" is a more widely accessible
statement of the existence requirement than "there must be a text
equivalent."  It is then necessary to state in additional refining clauses
what the text alternative should be like.   Trying to get there in one step
leaves too many people in the dust, not following the story.

Also, "definition of 'able to be expressed in words'" is the wrong tree to
be barking up, as the actual criteria for text alternatives are a compromise
between effectiveness (expressing as much information as possible) and
efficiency (not bogging the browsing user down in mindless detail without
an ability to control the terms of engagement such as verbosity controls or
hyperlink follow/ignore options).

The key is getting the questions to the authoring activity that cover the
waterfront of information that should be available and packaging it into
a structure so the user has enough control over speed vs. depth of information.

The "questions for authors" part is roughly

a) "Just tell them what you know."  Actually, this is going to be more than
survives, but it is the place to start.  The justification has to be a
justification for holding information back, not for including it.

b) Make separate evaluations of

(1) the thing by itself
(2) why this thing is here -- the contribution of this object to the 
ongoing story.

Prioritize based on your best guess as to what the user is most interested in
knowing as you partition the text into the following outline.

The proposal that I would bring to the WCAG group based on our experiences
with a succession of formats, is that all formats should provide the following
defined parts for text alternatives:

a) shortTitle -- must make sense in the absense of the media object
b) restOfTitle -- works as supplement to shortTitle or to media object playing
c) moreAdLib -- what you would put in a D-link or longdesc target.

  http://lists.w3.org/Archives/Public/w3c-wai-ua/2003JanMar/0054.html

The shortTitle would be a noun if the element has no interactive function and
a verb if it does.  As with the best-written ALT text of today.

The restOfTitle would be supplemental information on the order of what is
currently put in the html:any.title attribute and displayed as a transient
onMouseOver addition to the display.

The moreAdLib would be further hypertext content at the level of richness of
other text in the surrounding document.  This will be somewhat limited in
captions and other text targeted to 'tiny' delivery contexts.  But it should
not be more limited than the general limations imposed by the context.

I should probably be more explicit about the fact that I am thinking in
XML syntax.

<example
  class="pseudo XML">
  <element src=mediaReference>
   <title>
    <short>shortTitle</short>
    restOfTitle</title>
   moreAdLib
  </element>
</example>

This has not been cross-checked with the latest XHTML2.0 draft but it's close.

Al

>There's also no way to test equivalence where accessibility is concerned: 
>a person who's blind has no way to tell if a given chunk of text is or is 
>not equivalent to a given image, because that person doesn't have access 
>to the image and so can't perform the comparison.

The fact that the blind consumer can't repeat and verify the test does not
mean that it is un-testable.  Multiple disinterested parties evaluating the
relationship between text and non-text objects -- in isolation and in the
context of the enclosing document -- would provide a test that it is reasonable
for blind consumers to trust, and publishers to trust that blind consumers
would not be injured by following the results.

Al

[note]

As a matter of nerdliness, I do object to the claimed problems with the term
'equivalent.'  There are no philosophical or pragmatic problems with the term
'equivalent' until one tries to regard it as defining a guideline, a line
in the sand.

The concepts of "in effect" and 'enough' are implicit in the term "equivalent"
if one but checks with Webster.  This is why the word is used in the slippery
term "equivalent facilitation."

  http://www.m-w.com/cgi-bin/dictionary?book=Dictionary&va=equivalent

On the other hand "having a close enough effect" is a fine concept but an
unenforceably vague guideline.

So we do need a more concrete statement of how to assess what belongs in the
text option.


>If we push it hard enough, there is no such thing as a "text equivalent" 
>for *any* non-text element-- and if you ask people who do translation 
>professionally (especially but not only literary translation), there 
>aren't even "text equivalents" for *textual* elements.
>
>No one really expects a textual description of a painting or a symphony or 
>an actual (or fictional) event to be an exact equivalent for the lived 
>experience.  And it may be that thinking in terms of *description* 
>actually drives the text farther away from equivalence: a poem or a 
>symphony might come closer to providing an equivalent *experience* to that 
>afforded by a painting.
>
>There's also no way to test equivalence where accessibility is concerned: 
>a person who's blind has no way to tell if a given chunk of text is or is 
>not equivalent to a given image, because that person doesn't have access 
>to the image and so can't perform the comparison.
>
>But-- given that full equivalence is impossible-- it's still vitally 
>important to require meaningful descriptions/text equivalents for complex 
>images and other non-textual elements, including works of art.  That's for 
>the guidelines.  Techniques documents and examples will have to do the 
>work of modeling different ways of approaching the task, and we'll have to 
>get users who depend on those equivalents to give us some feedback about 
>which ones are most valuable.
>
>John
>
>
>
>John Slatin, Ph.D.
>Director, Institute for Technology & Learning
>University of Texas at Austin
>FAC 248C
>1 University Station G9600
>Austin, TX 78712
>ph 512-495-4288, f 512-495-4524
>email jslatin@mail.utexas.edu
>web <http://www.ital.utexas.edu/>http://www.ital.utexas.edu
>
>-----Original Message-----
>From: Gregg Vanderheiden [mailto:gv@wiscmail.wisc.edu]
>Sent: Friday, July 11, 2003 12:26 am
>To: w3c-wai-gl@w3.org
>Subject: RE: REF 1.1a - Add definition to 1.1 for ability to be expressed 
>in words
>
>Sorry Paul,
>
>
>
>I have to both agree and disagree with you on this one.
>
>
>
>Yes you can use words to at least partially describe things.  But they are 
>not equivalent unless you can use the words to reconstruct the original- 
>at least functionally.
>
>
>
>This is possible for many things but art and musical performances are not 
>among them.  (a score does not recreate a performance).
>
>
>
>We will have to work on this one carefully to not create a loophole but 
>also not create unattainable goals.
>
>
>
>Too tired to figure  this one out now.   But we need to do this very 
>carefully.
>
>
>
>
>Gregg
>
>  -- ------------------------------
>Gregg C Vanderheiden Ph.D.
>Professor - Ind. Engr. & BioMed Engr.
>Director - Trace R & D Center
>University of Wisconsin-Madison
>
>-----Original Message-----
>From: Paul Bohman [mailto:paulb@cpd2.usu.edu]
>Sent: Thursday, July 10, 2003 4:34 PM
>To: gv@trace.wisc.edu; w3c-wai-gl@w3.org
>Subject: RE: REF 1.1a - Add definition to 1.1 for ability to be expressed 
>in words
>
>
>
>I have never liked the phrase "that can be expresed in words."
>
>
>
>The truth is that anything can be expressed in words. Musicologists can 
>describe symphonies. Art critics can describe paintings. Even your 
>ordinary person can describe both of these. It may be true that the 
>description does not substitute for the experience of actually hearing a 
>symphony or of seeing a painting in person, but that is beside the point. 
>Anything can be expressed in words, no matter how inadequately.
>
>
>
>Like John, I don't wish to provide a loophole through which almost 
>anything can slip. Almost anything can be said to be impossible to express 
>in words if you mean that you want the reader to experience the 
>description in exactly the same way that the author does. I could argue 
>that it is completely impossible to give alt text to any image that would 
>truly substitute for not being able to see the image. No one can write 
>anything that would allow an individual who is blind from birth to be able 
>to visualize anything in exactly the same way that a sighted person can. 
>It simply can't be done. A person who has never heard a sound will never 
>experience music the way that a hearing person does, but you can always 
>describe music.
>
>
>
>In most cases, Web developers aren't going to post a link to a symphony 
>and say nothing about it. They usually have a reason for linking to it. 
>Maybe they want the listener to hear the difference between Barroque and 
>Impressionistic music. The differences can be explained in words. Maybe 
>the developer is just trying to sell CDs by giving sample music clips. The 
>selling points of the music can be explained. No matter what the purpose 
>is, it can be explained somehow.
>
>
>
>I would like to either drop the phrase "that can be expressed in words". 
>The important part of the checkpoint (making the function or information 
>available) is already expressed in the current wording (minus the 
>"expressed in words" phrase):
>
>
>
>"All non-text content has a text equivalent of the function or information 
>that the non-text content was intended to convey. [was 1.1]
>
>
>
>Paul Bohman
>Technology Coordinator
>WebAIM (Web Accessibility in Mind)
>www.webaim.org
>Center for Persons with Disabilities
>www.cpd.usu.edu
>Utah State University
>www.usu.edu
>
>-----Original Message-----
>From: w3c-wai-gl-request@w3.org [mailto:w3c-wai-gl-request@w3.org] On 
>Behalf Of Gregg Vanderheiden
>Sent: Wednesday, July 09, 2003 1:56 PM
>To: w3c-wai-gl@w3.org
>Subject: REF 1.1a - Add definition to 1.1 for ability to be expressed in words
>
>
>REF  1.1a  -   Add definition to 1.1 for ability to be expressed in words
>
>
>
>
>
>The phrase ability to be expressed in words is never defined.  Suggest 
>that in the definitions section, a new definition be added which would read:
>
>
>
>
>Ability to be expressed in words
>
>
>
>This refers to content that can be expressed accurately and unambiguously 
>in a reasonable number of words (for example, diagrams, charts, 
>illustrations, etc.)  Content such as a musical performance or visual 
>artwork is considered content that can not be expressed in words, since 
>this type of content relies heavily on the visual (or auditory) experience.
>
>

Received on Friday, 11 July 2003 12:09:30 UTC