W3C home > Mailing lists > Public > w3c-wai-au@w3.org > April to June 1999

Re: file name plus file size not a bad starting ALT

From: Charles McCathieNevile <charles@w3.org>
Date: Thu, 24 Jun 1999 10:41:07 -0400 (EDT)
To: Al Gilman <asgilman@iamdigex.net>
cc: w3c-wai-au@w3.org
Message-ID: <Pine.LNX.4.10.9906241026370.30718-100000@tux.w3.org>
As an alternative to not doing anything at all about ALT, filename and size
is useful. In addition, the file size (in particular) is valuable to people
who are in low-bandwidth situations.

But without some idea what the image is about, the information is still not
very helpful, because you don't have any way of deciding whether or not the
information is useful. For an example of how unhelpful this can be, point
lynx at http://www.footy99.com.au - something I do most weeks.

The argument seems to be the same as that for text only sites. It is an easy
way to automatically do a bad job. But you cannot be certain that you have
provided anything meaningful, let alone useful. To allow a tool to stop at
this seems to me a failure - it relies on any accessibility information being
included by an accident of nomenclature. As an initial default, there is a
risk that the benefit it provides will be outwieghed by it being interpreted
as an acceptable alt text in most contexts, and not requiring further author
attention. I think a tool which strongly encourages author involvement in
producing sensible and meaningful alteraative content, and which can maintain
metadata to offer more valuable defaults for alternatives where they have
been previously defined, will minimise that risk. A tool which does not
require any action on the author's part, or "hides" the ability to change
alternate text at the back of some chain of menus will maximise that risk,
which has been realised in a widespread pattern already.

There is another way to get the same information - the name of the resource
can often be determined by the User agent - which has it so it can ask for
the image, and the file size is one of the things HTTP sends with a head.

Charles McCN

On Thu, 24 Jun 1999, Al Gilman wrote:

  This issue came up tangentially on WebWatch and David asserted the AU
  guidelines would discourage use of (file name , file size) as a
  machine-generated initial value for ALT text.
  
  I am not sure if the benefits of this practice have been duly considered by
  the group, so I am repeating them here for your consideration and balancing
  off against all other relevant interests in this matter.
  
  Al
  
  At 08:48 AM 6/24/99 -0400, David Clark wrote:
  >Kelly,
  >
  >Just wanted to respond to this on behalf of CAST.
  >
  >You make a good point that, ideally, Bobby should look at the content of the
  >ALT text to try and detect whether it is just the file size. We have not
  >addressed it to this point because we have had other higher priorities. We
  >will be looking at it for future versions.
  
  Rules for checking HTML are being discussed and developed in the Evaluation
  and Repair working group in the WAI.  You can join.  See
  <http://www.w3.org/WAI/ER/> for more information.
  
  In particular, there is a working draft description of these rules at
  
  
  http://aprompt.snow.utoronto.ca/docs/Implementation.html
  
  Comments should be addressed to <w3c-wai-er-ig@w3.org> and follow-up to
  comments can be read at <http://lists.w3.org/Archives/Public/w3c-wai-er-ig/>.
  
  Yes, these people are cooperating with the author tools group.
  
  >
  >This problem also needs to be "nipped in the bud" by discouraging authoring
  >tool publishers from having this content be the default ALT text when
  >inserting images. The WAI Authoring Tools Working Group is including  this
  >recommendation in their guidelines.
  >
  
  This is debatable.
  
  I fear that the file size has gotten an unduly bad press because there are
  no representatives of the "low bandwidth network connection" user group
  involved in the discussion.  File name and file size suits the needs of
  this group of users very well.  For them one role of the ALT text is as a
  link label for the optional retrieval of the image.  As a starting
  insertion, if the user will take just a moment to edit based on this
  beinning, it makes a lot of sense.
  
  Al
  
  
  >Thanks,
  >
  >David M. Clark
  >CAST, Inc., 39 Cross St., Peabody, MA  01960
  >Tel 978-531-8555 x236 - Fax 978-531-0192
  >Email dmclark@cast.org
  >http://www.cast.org/bobby/
  >
  >
  >-----Original Message-----
  >From:	Kelly Ford [mailto:kford@teleport.com]
  >Sent:	Tuesday, June 15, 1999 8:46 PM
  >To:	webwatch
  >Subject:	[webwatch] Perhaps a Few Friendly Notes: http://www.miusa.org
  >
  >Hi All,
  >
  >One of my annoyances is the web page that sticks an alt tag of the filename
  >used for the graphic that includes not only the name but also the file
  >size.  Most of the time this information is meaningless but it tends to get
  >past validators and gets folks thinking that they've done their job.  I
  >mean do I really know that an alt tag reading "approved.gif (2614 bytes)"
  >means that a web page has been Bobby approved.
  >
  >Thankfully Bobby does ask, when running
  >
  >http://www.miusa.org
  >
  >through the service if the images convey more than what's contained in the
  >alternative text.  I'd like to see Bobby try and prompt for more than this
  >and perhaps watch for alt text that includes file extensions like.gif, jpg
  >and such.  Bobby could explicitly state that this probably isn't the most
  >desirable alt text to include.
  >
  >Anyway, if folks have a second you might drop a note to
  >
  >info@miusa.org
  >
  >asking this organization that deals with study abroad for folks with
  >disabilities to give a tad more attention to the home front.
  >
  
  >Kelly
  > 
  

--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, 24 June 1999 10:41:10 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:39:42 UTC