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

Proposed conclusion of null alt text debate

From: Leonard R. Kasday <kasday@acm.org>
Date: Fri, 25 Jun 1999 11:16:20 -0400
Message-Id: <>
To: w3c-wai-er-ig@w3.org
Like I mentioned, I'm going to be forwarding questions and comments on the
guidelines to the web content author guideline [WCAG] folks.  I'd like to
see how much we can wrap up the discussion of null "" and blank " " alt
text.  We don't think we need to come to consesus: WCAG has final say
anyway.  In fact, it's good to show both sides of difficult to resolve issues.

 So I'm going to list what I sense we agree on, and issues where we still
differ.  Once we satisfied that all points of view are represented, I'm
going to send it over.  Of course, if we do reach a consensus that would be
even better.

The following discussion runs through the tread started at


Here's my proposed summary. It includes some new things that I realized as
I wrote the summary: I've enclosed these in brackets [ ] to avoid sneaking
in something in this "summary".

 It includes the term null alt text to mean 

ALT=""    i.e. quote quote

and blank alt text to mean

ALT=" "  i.e. quote space quote.

There was a lot of discussion about Technique 5.6.1. 


 It was interpreted to mean that blank alt text " " is forbidden but null
alt text "" is allowed.  Several people pointed out that blank alt text is
not good HTML: browers are allowed to ignore leading and trailing space,
and in fact the HTML spec says that it "should" not be used.  However, it
was argued that these guidelines should only address things that impact

[However, 5.6.1 doesn't explicitly forbid blank alt text.  All it says is
that "Authors should not use spaces for the value of "alt" to prevents the
words from running together when the image is not loaded".  So first of
all, it only applies to this special case.  Also, the example is
alt="&nbsp;&nbsp;" which is an entirely different thing]. 

So the questions to WCAG are:

a. Is blank alt text " " forbidden always? If so  Why?
b. Is blank alt text " "  forbidden for purposes of spacing text apart?  If
so, why?
c. Is "&nbsp;&nbsp;" forbidden always? If so, Why?
d. Is "&nbsp;&nbsp;" forbidden for purposes of spacing text apart? If so, Why?

Another suggestion was that blank alt text be used for spacers " " because
that coveys the meaning of a space.  Objections to this were the "bad HTML"
argument used above, and also that it contradicts 5.6.1.

The following applies to checkpoint 13.1 which requires meaningful text in
a link

Suggestion to WCAG:  allow null alt text in links in certain circumstances,
per following rules:

    If the image is the only item in the link
    and there is no other usable link to that item on the page,
    then neither null or blank alt text is allowed.
    for example,

    <A href="foo.html">  <img alt="" src=...> </A> 
    is illegal if no other link on the page points to foo.html.

    However, if usable text or image with meaningful alt text is
    included in the link, then null alt text is allowed.  E.g. the 
    following are legal.

    <A href="foo.html">  <img alt="" src=...> prices </A> 
    <A href="foo.html">  <img alt="prices" ...><img alt="" src=...> </A> 

There's another situation on which so far at least we didn't reach
consensus: the case where an image and text which are visually adjacent
point to the same url.  Examples are images which are bullets or images, as
in www.whitehouse.gov. This is very common on the web.  Sometimes these
pages can be rewritten to give a similar visual appearance with image and
text as part of the same link, but this may not always be possible, e.g. if
tables are used for layout.

It was suggested that null alt text be allowed in such cases for the image
to avoid tiresome redundancy where every link gets spoken twice.  On the
other hand, it was argued that some browsers would reveal that there's a
link there, e.g. lynx with certain options such as lynx -number_links
-hiddenlinks=merge [or using the l command].  [possibly also browsers like
pwwebspeak or home page reader] It was argued that a user would lose
confidence in the pages accessibility to hear links without ALT text.  

On the other hand, [under some circumstances] the links would be invisible,
and could  can be switched off at users discretion, in some cases by
turning off speaking pronunciation [or by turning off the lynx options
mentioned above].

Question to WCAG: to allow or not allow null alt text for this circumstance.

A final issue to WGAG

In the deprecaed [put presumably allowable] example in 5.6.1, "spacer" is
used for alt text of a spacer element. There was objection this since it
was felt to just clutter up speech  [and contractions the recommendation of
using null alt text earlier in 5.6.1]

I'm going to be away from my email till thursday (will be at RESNA).  I
want to give people a chance to respond to this before claiming it's our
final comments. So I'll just send a heads up to WCAG with a pointer to the
archived mail, and send the "official" one next week when I get back,
updating with any comments.

I will try to fold in any comments that come today into a revised version
of this to which I can point WCAG.

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)
Received on Friday, 25 June 1999 11:14:01 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:29:59 UTC