Re: Study project: ALTifier or text-equiv -- Please choose!

Hi,

At 17:46 26-10-98 -0800, Victor Tsaran wrote:
>I think, this is a good argument as well. If we start to rely on
>automatic conversion tools, web designers will abandon accessibility
>problems at all.

Maybe you are right. But experts think that 99% of the webpages on the
Internet have accessibility problems. I'm very active in writing messages
to webmasters asking them to make their website more accessible. I give
them information how they can do it. 
Most of these webmasters say: yes I shall add ALT tags. But a lot of
webpages here in the Netherlands include image maps and the webmasters
don't know how they can make these image maps accessible without changing
the look of their pages. 
To make webpages accessible seems never to have any priority. The website
of our parlement at http://www.parlement.nl has a lot of missing alt tags
and include a lot of image maps. These image maps are not more than a menu.
I asked the webmaster why not create a text menu instead of that image map.
I asked them also what is most important the concept of the website or
accessibility of the parlement website. The want to discuss the
accessibility problem in their project team. Maybe they are going to
discuss this for a year in their project team. In the same time hundreds of
new websites come on the Internet with the same problems. 
I get tired to send messages to all these webmasters every day. To discuss
that it is important to make a website accessible. 
I don't give up, but at the sohort time it makes the accessibility of hte
Internet not better. 
So for the short term a tool that the end-user can use to make
text-equivalent of a graphical page has in my opinion priority. 
In theory it was best if such a tool uses all techniques described at
http://www.w3.org/WAI/ER/text-equiv.htm. Beginning with the more simple
technique and if this does not solve the accessibility problem use more
advanced techniques.
Most accessibility problems are caused by clickable images without ALT tag
and image maps. A text-equiv tool should first look for ALT tags. If there
are no ALT tags:

<A HREF=url><IMG SRC=url></A> 

The best text seems the GET <TITLE> of url to me.

<MAP> <AREA HREF=url> ... 

If there is no ALT of AREA. This will be the case most time because this is
HTML 4 and not TITLE of AREA, 
GET <TITLE> of URL has also priority here.

<IMG SRC=url> 

If there are no ALT of IMG and no TITLE of IMG I think the priority is
Filename part of url.
This is not a solution if the image presents a  lot of text, but I think
using OCR techniques make the tool very complex and can make the tool slow.

<FRAME SRC=url> 

If there is no TITLE of FRAME, GET <TITLE> of URL has priority.

<A HREF=url> "click here" </A> 

If there is no TITLE of A, GET <TITLE> of URL has priority. 

<INPUT SRC=url> no ALT 

If there is no TITLE of INPUT, NAME or ID of IMPUT has priority.

Question?

I personaly prefer a text-equiv tool that I can install like a plugin in my
browser. But this is a problem for Lynx and Net-Tamer. What options are there?

Regards Peter Verhoeven
Internet : http://www.plex.nl/~pverhoe (The Screen Magnifiers Homepage)




 At least if designers would give meaningful names to
>their images, like help.gif or options.gif, one could try to guess the
>content of the image by cutting off the extion of the file. I use this
>technique  lot in an ttempt to guess the destination of the image on
>the page.
>Best regards,
>Victor
>
>
>
>
>
>---"Leonard R. Kasday" <kasday@acm.org> wrote:
>>
>> Al Gilman wrote:
>> 
>> >So I see application for the text-equiv techniques on both the
>> >author and user side, and I see the author scenario encompassing
>> >both automatic suggestions and manual judgement and rewording.
>> 
>> LRK::
>> 
>> Well, logically I have to agree with that, but, to reveal a personal
>bias
>> here, are there any cases where the automatic text would be as good
>as text
>> the author wrote him or herself, aside from 
>> 
>> - copying text link to a ALT text for image or image map area with
>same URL
>> - and visa versa of the above
>> - all-text image read by OCR (tho even there an error
>>   could produce a real knee slapper)
>> 
>> I'm worried that authors would get into habit of using the
>automatically
>> generated text in cases where it's inferior to what they could
>easily do
>> themselves.  
>> 
>> Len
>> -------
>> Leonard R. Kasday
>> Institute on Disabilities/UAP at Temple University, Philadelphia PA
>> email:     kasday@acm.org
>> telephone: (215} 204 2247
>> 
>> 
>
>==
>Hi, visit me at:
>http://tsarnet.home.ml.org
>
>_________________________________________________________
>DO YOU YAHOO!?
>Get your free @yahoo.com address at http://mail.yahoo.com
>
>

Received on Tuesday, 27 October 1998 03:40:31 UTC