W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > January to March 1998

Re: [w3c-wai-ig] <none> (fwd)

From: nir dagan <dagan@upf.es>
Date: Mon, 16 Feb 98 19:43:16 MET
Message-Id: <199802161843.TAA25325@sahara.upf.es>
To: w3c-wai-ig@w3.org

In my view the best approach is to assume first that the Specs
are implemented correctly and to build the guidelines (as 
a first iteration) based on that.

Many of the problems you mention are dealing with the controversial 
implementation of ALT in graphical browsers.

It seems that the designers of these browsers assume that users are 
people who can (physically) view the images but sometimes choose not do so,
e.g., shutting off images to  save download time.

For this reason the boxes of images are drawn allowing the users to download
images separatly if they so choose.

I think that the specs are very clear about the new OBJECT element, more 
than about alt to images. The spec says quite explicitly that if images are 
turned off the content of the OBJECT element must be shown. No box or cut off
alt text etc. 

I noticed that when I used the OBJECT element with an icon-link in one of 
my pages (with Height and Width specified) MSIE4.0 for Windows 3.1 
implemented
it wrong: Providing scroll bars and hiding the little image behind them,
and the link didn't work. This was with images turned on.
(I took it off my page...)

A more careful study of the implementation of OBJECT should be done.
(I'll write some test pages when I have the time...)
 
But in principle using OBJECT when the text is essensial seems to be the 
solution, if OBJECT is implemented as the specs say.
(and maybe putting decorative image with OBJECT with no content)

Nir Dagan.

email: dagan@upf.es
URL: http://www.econ.upf.es/%7Edagan/ 

I have some more short comments below:

> 
> I have been trying to look at the different postings on the ALT Text issue 
> and dig out the common threads and see if there is a bottom line or simple 
> solution.
> A few interesting points however, seem to complicate this whole discussion. 
>  The answers seem simple if we just forget some of them.  But they are all 
> real, and when I try to account for all of them - it is a little tougher. 
>   Let me list what I have as a starting point to getting a complete list.
> 
> 
> 
> 1)  Unlike text browsers, graphic browsers currently leave an image on the
> page, even if there is a null alt-text.  Thus, people using graphic screen
> readers are still left with the screen reader announcing "image" without
> any explanation as to what the image is.  Thus, marking decorative images
> with no alt-text or with alt="" are identical.

True. but no-alt is invalid. N.D.

> 
> 2)  What some people consider decorative other people consider information.
>  More importantly, if alt="" is a legitimate strategy for anything that
> does not see important, it's possible that page authors might use it
> overmuch, removing much content.

I have seen many pages myself with *links*  alt tagged with an 
empty string.  The author here just tried to get "formal validity"
of his page (e.g. the HTML writers guild) rather than using his head.

In principle the author should be the one who decides what he wants to show
to users who don't see the image.  This principle also
resolves the function/description debate. The author should think
 what does best for his readers  N.D.


> 
> 3) Alt-text was originally designed to provide the same function as the
> graphic.  When the graphic is a link, the alt-text would generally just
> provide text to indicate what the link was (not what the graphic looked
> like).  However, when a graphic is not used as a link, then the purpose of
> the graphic is usually to present something graphic, and the alt-text would
> in that case have to describe the graphic.  This has led to a lot of
> confusion.
> 
> 4) Now we have a "title" whose function I have sometimes seen depicted as
> providing additional information about the function and sometimes as
> providing additional information about the appearance.  Again, if something
> is graphic, its function is its appearance.
> 
> 5) Finally, we have objects that need alternative text associated with
> them, but for which alt-text is not supported.  Thus, "title" takes over
> the role.
> 
> So it appears to me that we have somewhat of a mess that has three major
> characteristics:
> 
> a)  Some users who are blind want information about graphics, including
> decorative graphics, while other users do not.  Thus, whether to make
> graphics invisible has no single right answer.
> 
> b) There seems to be much confusion, and not much consensus, about the
> exact role for alt-text and title on different objects, and on the same
> object used in different ways (e.g., a graphic).

I think that a great deal of the problem is that title attribute is not
well defined in the specs. We don't know what to expect with it.  N.D.


> 
> c)  Title appears in places where alt-text does not, making the division of
> roles between alt-text and title more complicated.
> 
> 
> Right now I'm scratching my head while reading all the messages
> going by, and trying to see whether there is a strategy we might use for 
> sorting all of
> this out.
> 
> It obviously needs to be backward-compatible,
> 
> and it would be a shame if it turned out we needed to define anything 
> differently than we just did in HTML 4.0.  (Like the role of alt text vs 
> title or whatever)
> 
> However, above all, we have to think of something that is more
> straightforward and easy to explain than what we have now.  It's quite
> evident that what we have right now is abundantly confusing even to people
> who are advocates and well-versed in the area.
> 
> Perhaps instead of restating positions and opinions about what things stand
> for or have stood for, etc., we should focus on trying to summarize what it
> is that we're trying to achieve with the combination of alt-text and title
> (and for that matter long-desc), then list the constraints that must be met
> by a solution.  Then finally we should try to propose something that is
> simple, straightforward, consistent, and provides maximum accessibility
> across the different types of objects today and whatever we can predict
> about the future.
> 
> 
> 
> I think the eventual strategy needs to handle all of the following issues
> together in a comprehensive solution strategy:
> 
> I-1) The role of alt-text (on images used as anchors, stand-alone images, 
> and
> anywhere else it's supported);
> 
> I-2) The role of title (for same conditions as above, as well as for 
> objects
> where alt-text is not supported);
> 
> I-3) How to handle decorative graphics and spacer graphics;
> 
> I-4) Proper use of long-desc (and temporarily D-tags?)
> 
> 
> 
> In a separate post, I think we will pull out just the items from this 
> message and what we've seen in other messages that fall in the categories 
> I've described above, so that we can begin to build a list of requirements, 
> concerns, etc., and see if we can get all the issues and constraints down 
> in one place, so that we can then work forward to an internally consistent 
> and straightforward solution.
> 
> 
> I also wonder if we should take this back to the GL list and work on it, 
>  then bring it back to the IG  and other lists for discussion.
> 
> 
> 
> 
> Gregg
> 
> 
> 
> 
> -- ------------------------------
> Gregg C Vanderheiden Ph.D.
> Professor - Human Factors
> Dept of Ind. Engr. - U of Wis.
> Director - Trace R & D Center
> gv@trace.wisc.edu    http://trace.wisc.edu
> FAX 608/262-8848
> For a list of our listserves send "lists" to listproc@trace.wisc.edu
> 
> 
> 
Received on Monday, 16 February 1998 13:50:13 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 19 July 2011 18:13:39 GMT