W3C home > Mailing lists > Public > w3c-wai-er-ig@w3.org > June 1999

RE: Guideline 1 in The evaluation techniques document

From: Bruce Bailey <bbailey@clark.net>
Date: Tue, 22 Jun 1999 15:46:43 -0400
Message-ID: <01BEBCC6.D1C44540.bbailey@clark.net>
To: "w3c-wai-er-ig@w3.org" <w3c-wai-er-ig@w3.org>
I agree that we must, at all costs, stay in synce with the WAI guidelines. 
 I will not, however, refrain from complaining.  ALT=" " is much closer to 
function (for spacer images) than ALT="spacer" which is, obviously, 
descriptive.  I wish the debate over the validity of ALT=" " were archived 
someplace.  My understanding is that the final opionion on this matter was 
reached because the HTML specs disregard all leading and trailing white 
space.  This means that ALT=" " should be treated as ALT="" and that a 
paradox is produced by having a string to produce the NULL value!

ALT="" is used for LOTS more things than spacer images.  One technique that 
I think is fairly common is to divide a large graphic into pieces so that 
it loads faster.  ALT text is really only required for the first piece and 
ALT="" would be *prefered* for subsequent pieces of the same graphic. 
 Check out http://www.apple.com/ for an example.

The rule that you are trying to enforce is:
> Provide text equivalents for all images, including invisible or 
transparent images.
> If content developers cannot use style sheets and must use invisible or 
transparent images
> (e.g., with IMG)  to lay out images on the page, they should specify 
alt="" for them.

The prohibition against ALT=" " occurs ONLY as part of a depreciated 
example and it is NOT explained.  We are NOT talking about writing an HTML 
validator.  What is *OUR* (the WCAG) *REASON* for prohibiting ALT=" "? 
 BOTH of deprecated examples at 
http://www.w3.org/TR/WAI-WEBCONTENT-TECHS/#spacer-images are meant to show 
the *wrong* way to do things.  ALT="spacer" is *NOT* suggested by the 
techniques document.

Until we get an definitive answer to this, I propose the following:

- ALT=" " is allowed.  (No warning, nothing.  Not untill we can say (with a 
clear reference) with authority WHY this is not allowed.

- ALT="" is suspicious and, if found, the user is prompted with "ALT text 
for this image is NULL.  This is acceptable only if this image is used 
solely for decorative purposes and is completely free of meaningfull 
content, for example, as a spacer image."

ALT="" should NOT be allowed when the image is the only item in a link. 
 For example:
<a href="foo.html><img src=foo.gif alt=""></a>

This would mean that our authoring tool would have to be dynamic enough to 
<a href="foo.html><img src=foo.gif alt="">go to foo</a>
but then generate an error condition if "go to foo" were deleted!

This is tricky because the "easy out" is to never allow ALT="" -- which IS 
in conflict with the WCAG techniques document!

Rules enforced by authoring agents CAN be more strict than the WCAG.  This 
would alow us to include the ALT="" exclusion!

I am all for dropping any references to the "TITLE" attribute.  I raise the 
issue to make the point of how the use of TITLE and ALT(ernate) are 


On Tuesday, June 22, 1999 11:03 AM, Chris Ridpath 
[SMTP:chris.ridpath@utoronto.ca] wrote:
> Bruce,
> We are trying to stay in sync with the WAI guidelines so to follow
> http://www.w3.org/TR/WAI-WEBCONTENT-TECHS/#spacer-images:
> ALT="   " (1 or  more spaces) is NOT allowed.
> ALT="" (NULL) IS allowed for spacer images.
> The WAI techniques also suggest that the ALT text for spacer images be
> "spacer". I like this better than the NULL text. I propose that I modify
> technique 1.1.A so that NULL ALT text is suspicious and if found, the 
> is prompted with: "ALT text for this image is NULL. If the image is used 
> spacing use 'spacer' as the ALT text."
> >I disagree (rather vigorously) with the suggested text for bullets and
> >horizontal rules.  ALT (alternate) should replace function and not form.
> >
> >* 2. If the width of the image and the height of the image are both less
> >than 20 pixels, suggested text should be "bullet".
> >
> >Suggested text should be "-" or "." or "*" or "+"!
> >
> >* 3. If the height of the image is less than 20 pixels and the width of 
> >image is greater than 150 pixels, suggested text should be "horizontal
> >rule".
> >
> >Suggested text should be "----"!
> >
> That sounds fine to me. It makes it more language independant.
> >We could recommend that authors include the "TITLE" attribute (which
> >should, if included, be "bullet" or "horizontal rule").  It is not clear 
> >me that any browsers support the use of TITLE for images though...
> >
> This may be added complexity for something that is not used. Can we drop
> this recomendation?
> Chris
Received on Tuesday, 22 June 1999 15:48:04 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:01:28 UTC