W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > July to September 1999

Re: How to describe Flowcharts, Schematics, etc

From: Charles McCathieNevile <charles@w3.org>
Date: Thu, 12 Aug 1999 18:20:27 -0400 (EDT)
To: "Leonard R. Kasday" <kasday@ACM.org>
cc: w3c-wai-ig@w3.org
Message-ID: <Pine.LNX.4.10.9908121816560.10832-100000@tux.w3.org>
It has been suggested that we should be asking SVG to provide the necessary
semantics to connect objects by a vector, as well as connecting points in
space, since it would then become relatively simple to transform this into
text, at least as a thought experiment.

But maybe we should look at hot to describe the relationships between
elements of a diagram, and go from tehre to representing the elements and
relationships in SVG...

I think we really want to be able to go both ways. The nice thing about SVG
as a fframework for thinking about the idea is that it enables all kinds of
neat stuff like defining and describing a component of an image, and then
using it as an atom, in a recursive process that works like complex programs.

(The difficulty of describing how complex SVG diagrams work is probably
similar too. Sigh)

Charles McCN

On Thu, 12 Aug 1999, Leonard R. Kasday wrote:

  Steven McCaffrey had some suggestions on describing flowcharts, schematics,
  etc that we chatted about offline and I wanted to bring back to the list.
  Hs proposal is that ALT text or even LONG DESCRIPTION is not always enough:
  interactivity is needed.
  The thread is as follows:
  >Hi Len:
  >That's an excellent question and well worth bringing up at this time.
  >Since I am totally blind myself and use such technology and have seen such
  diagrams in the past during my undergrad days (esp. data structure diagrams
  in computer science classes)  when I had sight, I may have some overall
  >Whenever I have discussed this topic with sighted individuals familiar
  with such diagrams, my question is really a simple one.  What information
  are you seeking via the diagram that you could not get as easily or at all
  without it?  It's often surprising that I usually get a long pause in
  >I try to clarify by saying "What questions are you asking yourself in your
  mind when you look at the diagram?  What information is conveyed by knowing
  relative (i.e. relational )position (e.g. "at a higher/lower  level than"
  or "is a sibbling of" on")"  How do you find the answers to these questions
  with your eyes (i.e. describe the algorithm you are, perhaps unconsciously
  so perhaps difficult, using).  The computer code behind such diagrams can
  do the same thing.  
  >So, that's it in a nutshell.  Ask yourself what questions you put to the
  diagram.  Ask how you find the answers with your eyes.  I would design the
  page with questions as links.  That's my half-baked idea at the moment.
  >> "Leonard R. Kasday" <kasday@acm.org> 08/11 5:29 PM >>>
  Thanks for the suggestions Steve.
  What you're suggesting sounds like more than just ALT text or even a long
  description.  It sounds like something that needs to be interactive.  
  Could be done with links as you suggest.  In fact, I tried something like
  that for table access.  It listed table cells and by choosing links you
  could navigate by row or column and announce where you were.  It's at
  http://astro.temple.edu/~kasday/web_access/table_access/ if you're curious.
  Could modify this to walk a tree for example.
  It would be better I think to always have defined keys to do the actions.
  Actually, this could be kluged by using ACCESSKEY and jumping to new pages,
  but would be quite slow.  Best to be interactive at the client level.
  Hmmm.  Might work with javascript if it changed the whole page and the
  screenreader could follow it.
  On  Thu, 12 Aug 1999 07:42:38 "Steven McCaffrey" wrote:
  Hi Len:
  You are quite welcome and thank you for bringing this under discussed topic
  to a wider audience.  Yes, I do think it needs to be highly interactive at
  the client level.  Tree traversal was exactly what I was thinking. Since,
  of course, all particular instances of tree structures can be represented
  by the same, what used to be called, "abstract data type", my idea was to
  have subject-specific terminology correspond to the abstract operation.  To
  take a very simple example, if I have an organization chart, I might ask
  "Who is the director of the organization?"  or "Who is the head of my
  department/division?" or "Who is my counterpart in office x?"etc.  These
  are just some generic questions I can think of at the moment.  The
  particular questions chosen must come from the person(s) who put the chart
  out on the web.  What information do they want the viewers of that chart to
  obtain?  These are just ordinary information design considerations applied
  to the specific question of what any given diagram/chart is for.  As I
  remarked earlier, this is a little more difficult, because in many cases,
  the knowledge of what information is conveyed by a diagram/chart is, I
  think, often on an subconscious level and probably someone trained in
  information engineering might be able to help turn implicit questions into
  explicit ones.  However, I think anyone can do this with a little
  Your table structure sounds interesting and I'll check it out soon.  
  Sure, feel free to repost my response to the list.  I am never quite sure
  how much depth to post to a list on this topic.  It happens to be both very
  important and interesting to me but I have not seen a wide discussion on
  this to date.  I would like to see this discussed in some depth since I
  think this kind of material is going to be more ubiquitous soon and is not
  easily transformed into an accessible format.
  Steven McCaffrey
  Information Technology Services
  Leonard R. Kasday, Ph.D.
  Universal Design Engineer, Institute on Disabilities/UAP, and
  Adjunct Professor, Electrical Engineering
  Temple University
  Ritter Hall Annex, Room 423, Philadelphia, PA 19122
  (215) 204-2247 (voice)
  (800) 750-7428 (TTY)

--Charles McCathieNevile            mailto:charles@w3.org
phone: +1 617 258 0992   http://www.w3.org/People/Charles
W3C Web Accessibility Initiative    http://www.w3.org/WAI
MIT/LCS  -  545 Technology sq., Cambridge MA, 02139,  USA
Received on Thursday, 12 August 1999 18:27:01 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 13 October 2015 16:21:05 UTC