W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > October to December 1999

Re: proposal for "null" alt-text (was "Re: A few thoughts...")

From: Bruce Bailey <bbailey@clark.net>
Date: Sat, 20 Nov 1999 09:24:12 -0500
Message-ID: <3836AEDA.9C860B25@clark.net>
To: w3c-wai-ig@w3.org
CC: GARETH P PARKINSON <298gpp@tay.ac.uk>
The difference between null and missing ALT text is the real life behaviors of
browsers.  In somecases, MISSING alt text is PREFERABLE to null alt text!

>From the User Agent Accessibility Guidelines (now in "last call" status) at
URL:
http://www.w3.org/TR/1999/WD-WAI-USERAGENT-19991105/#gl-content-access
<Q>2.8 When alternative text has been specified explicitly as empty (i.e., an
empty string), render nothing.</Q>

This is how Lynx currently behaves, but not pwWebSpeak.  The guidelines say
nothing about what to do about missing alt (and  invalid html, of which would
be one example.  Lynx (and the big two with image loading off) generate
something like [IMAGE].  This is different behavior than ALT=" " or
ALT="nbsp;".  The bigger problem is when an image is a link.  With Lynx, each
link normally gets a number, so the following snippets of code are shown next
to the default rendering with Lynx:
<A HREF="foobar.html"><IMG SRC="foobar.gif"></A>    [1][LINK]
<A HREF="foobar.html"><IMG SRC="foobar.gif" ALT="Foobar"></A>    [2]Foobar
<A HREF="foobar.html"><IMG SRC="foobar.gif" ALT=" "></A>    [3]
<A HREF="foobar.html"><IMG SRC="foobar.gif" ALT=""></A>

Notice that the last example generates nothing!  If "hidden" links are the ONLY
links on a page, Lynx generates a warning, but otherwise these hidden links are
effectively invisible.  (One can list all links on a page, or look at the
source code, but these are extreme time consuming measures.)

We have been over this before, and many have made the arguement that ALT="" is
frequently appropriate.  I agree that it should be valid html, but I think this
is a point where the WCAG could make a clear unabigious distinction.

The problem with ALT="" is that it is ambiguous.  You don't know WHY the author
put that in there.  Is he just trying to get his code validated?  Is this
REALLY just a decorative, content-free graphic?  Does he have something against
text-based browsers?  Is this some kind of spacer image?  Did he just forget?

IMHO, ALT="" is NEVER prefereable to other choices.

For decorative graphics, use a single non-alphanumeric character.  There is
ALWAYS a punctuation mark that is closer in meaning (closer than nothing
anyway) to a "content free" image.

The burden of repeated links (one big arguement for using ALT="" within <A HREF
... /A>) is much less of a problem than the chance of missing content and
links!

Bruce Bailey


GARETH P PARKINSON wrote:

> Hi Len,
> I was wondering what effect placing a null value in the alt tag. Doesn't
> this result in the same problem as having no alt tags at all, in that the
> user will be prevented from finding out the reason for the image? I would
> assume that in most cases the image has some meaning, and therefore would
> benefit the user if they could have a description of it.
>
> I don't doubt that on a very busy interface there might be call to reduce
> the amount of text (although perhaps the page is overloaded anyway?) , but
> such decisions should be made with a lot of caution, and only when there's
> no other way to simplify the page. Otherwise,it might wrongly encourage
> designers to use null values on images that would enhance the page.
>
> Gareth Parkinson
>
> -----Original Message-----
> From: Leonard R. Kasday <kasday@acm.org>
> To: Scott Luebking <phoenixl@netcom.com>; <w3c-wai-ig@w3.org>
> Date: 15 November 1999 16:09
> Subject: Re: A few thoughts on using dynamic web pages to improve
>
> >Hi Scott,
> >
> >Thank for the extra example. I just want to check one thing to preempt any
> >misunderstanding.  Even though you offer a page with alternate layout and
> >other changes to make browsing more efficient with screenreaders, you still
> >make the default page "accessible" in the conventional sense, right?
> >
> >For example, the default page would still satisfy the web accessibility
> >guidelines... alt text, table reading order, etc.   It's just that the
> >alternate page would be more efficient to use.
> >
> >As for the details...
> >
> >1. It does seem useful to have an alternate layout where the section of
> >links and references come at the end instead of the beginning.  It might
> >also be useful to have a link at the start of the page to jump to that
> section
> >
> >2. You speak of leaving out decorative images, but why not accomplish that
> >by simply having null alt text?  A ha.
Received on Saturday, 20 November 1999 09:24:41 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 19 July 2011 18:13:45 GMT