W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > January to March 2002

RE: Creating accessible tables for layout and data: alt attributes

From: Jason Megginson <jason@bartsite.com>
Date: Thu, 31 Jan 2002 11:00:29 -0500
To: "'Joel Ward'" <ward_joel@bah.com>, "'WAI List'" <w3c-wai-ig@w3.org>
Message-ID: <001001c1aa70$68ba8d80$fe1ba8c0@jmegginson>
If the background image is within a table, is it ok to use a summary
attribute to convey the information about the image along with the
purpose of the table?  

Sorry to bring this subject up again but Joel makes a good point. 

-----Original Message-----
From: w3c-wai-ig-request@w3.org [mailto:w3c-wai-ig-request@w3.org] On
Behalf Of Joel Ward
Sent: Thursday, January 31, 2002 10:36 AM
To: WAI List
Subject: Re: Creating accessible tables for layout and data: alt

Hi Everybody,

I thought I witnessed a recent discussion here that was very similar.
thread was about background images (Subject: background-image in CSS).

What I learned from that discussion was the need for us, as web
and developers, to effectively share information about all of our
content to
all users.  This means that if an image adds something to the page,
users of
assistive technology might find that information useful, too.

I appreciated the example about a deaf person at a lecture.  Summary in
nutshell: Since the interpreter did not convey the sounds of a fire
the deaf person had no idea why the lecture had paused and what was
on.  The interpreter didn't feel the information was useful, but the
person did.

Not adding alt text to all images (besides spacers, of course) is along
those same lines.

Of course, that example was about important information, not decorative
information.  But the argument can be made that the user should be the
to filter what is important and what is not important.

If that image of a tree means something to a visual user, why can't it
something to a blind user?  And if it's not important enough to warrant
text, why include it at all?

There is that fine line between what is important/useful and what is
frivolous.  A very subjective line.


----- Original Message -----
From: "Jouni Heikniemi" <jth@dns.mikrobitti.fi>
To: "Jim Byrne" <j.byrne@gcal.ac.uk>
Cc: "W3c_Access" <w3c-wai-ig@w3.org>
Sent: Thursday, January 31, 2002 5:52 AM
Subject: Re: Creating accessible tables for layout and data: alt

> On Thu, 31 Jan 2002, Jim Byrne wrote:
> > I like to provide as much text on my pages as possible, just in case
> > search engine robot comes along and decides to add my page to a
> > database.
> I refrain from commenting your decision on the alt text issue, since
> obviously now know the practices and common suggestions made by the
> accessibility society - and have made up your mind accordingly. Your
> your rules. However, I'd like to make a point about the paragraph
> above.
> You shouldn't aim at getting high search engine index ratings by using
> texts which have no real relation to the content of the page. If I
> _really_ want to know something about, say, "pink flowers", I wouldn't
> want Google to return a set of web accessibility pages simply because
> have pink flowers as their decoration.
> However, this issue is debatable since many search engines support
> search which do index the images by their alt texts (among other
> Therefore, someone searching for an image with pink flowers might
> benefit from your "wrong" use of alt texts.
> My suggestion is that you promote accessibility and generic text-mode
> readability by leaving the alt texts empty for the decorations, but
> produce a separate gallery page ("images by me" or something similar)
> with appropriate titles and alts for those who are specifically
> for images. The search engines will index that one just as well.
> If you insist on leaving the alt texts into the text, you might
> using [ ] (alt="[ decorative photo of pink flowers ]") around the
> Though this could have adverse effects with speech synthesis, it will
> pure text-mode browsing by clearly pointing out what is alt and what
> normal text.
> Jouni
Received on Thursday, 31 January 2002 10:58:24 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:36:07 UTC