WebAIM - Web Accessibility In Mind

E-mail List Archives

Thread: A larger discussion (was RE: Inline Images and ALT text)

for

Number of posts in this thread: 24 (In chronological order)

From: John Foliot
Date: Wed, Jan 14 2009 12:30PM
Subject: A larger discussion (was RE: Inline Images and ALT text)
No previous message | Next message →

Patrick H. Lauke wrote:
>
> > Contrast: "Jared Smith, M.S., is the Associate Director for WebAIM.
> > Jared
> > Smith" vs. "Jared Smith, M.S., is the Associate Director for WebAIM.
> > [Photo - Jared Smith]" Most screen readers, unless configured for
> > maximum
> > verbosity, tend to ignore/not voice the square brackets, and so would
> > say
> > aloud: "Jared Smith, M.S., is the Associate Director for WebAIM. Image -
> > Photo - Jared Smith".
>
> In this case, even better: "Jared Smith, M.S., is the Associate Director
> for WebAIM." and a nice alt="" in the image?
>

...and in fact, this is what Jared is doing on his site now
[http://tinyurl.com/7cm9d9]. (P.S. Jared, I'm/we're not singling you out for
any particular reason...)


<for discussion>

The inclusion of a photo of Jared is done deliberately. It is not a
"decorative" image per-se, it is provided to allow folks to know what Jared
looks like, so that (for example) when you run into him at a conference you
can say hi to him, or whatever. As such, it is an "image of value" as
opposed to something that is purely decorative in nature. It is also
entirely feasible that a non-sighted user might want a copy of that image
(for a report, to add to their social network page, to share with an
associate, etc.), yet by not indicating that the image is part of the page
(and using alt=""), we/you are deliberately "hiding" that information "of
value" from the non-sighted user.

I posit that this is wrong.

This relates to a similar situation I encountered earlier this month here at
work: our central library is working on a next generation search tool for
the collections they maintain (and it is one of the largest private
libraries in the Western U.S., roughly 6.5 million titles). The results
page is offering an image of the book cover as part of the results. What to
do? Is alt="" the answer?

Using the same argument from above, I suggest no - there is both a reason
and a value in providing that image to the sighted user, and to not do so to
the non-sighted is unfair, unbalanced, and may in fact be a point of
"liability" (Note: I AM NOT A LAWYER!). Yet, at the opposite end of the
discussion, it is simply not feasible to provide a proper alt value for all
of the book covers in the databases (complicated by the fact that at least
one of the "image" db's is outside of our control - it's a Google's), and
echoing back alt="<?php echo($Booktitle); ?>" introduces a certain amount of
redundancy, since preceding the image is <h2><?php echo($Booktitle);
?>"</h2>.

At this time, I have suggested a middle-ground solution (alt="Book cover"),
fully aware that if the user is researching 30 books, that too will become
tiresome. However it is a simple fix for the developers (it's static text),
and it's better (IMHO) than "saying" nothing. I queried a few daily AT
users that I know and asked their opinions; that very small feedback further
suggested that what I was proposing was an acceptable compromise, but a
survey panel of 3 is hardly definitive.

So what do others think?

</for discussion>

JF
============================
John Foliot
Program Manager
Stanford Online Accessibility Program
http://soap.stanford.edu
Stanford University
Tel: 650-862-4603

Soap Is a program directed by the
Vice Provost for Student Affairs
============================

From: Randall Pope
Date: Wed, Jan 14 2009 12:45PM
Subject: Re: Inline Images and ALT text
← Previous message | Next message →

Hi all,

This brings another interesting point. Normally I usually caption the
photos that describe the picture in one or two sentences then added an
enlarged photo for the low vision people who cannot see the regular size
(well this is what the audience wants). So my question is: does a good
descriptive caption take the place of the alt text? In other words can I
just type in the double quote in the alt text if I have a good descriptive
caption next to the photo?

Many thanks for your help.

With Warm Regards,
Randall "Randy" Pope
American Association of the Deaf-Blind
Website: http://www.aadb.org

301 495-4402 VP/TTY
301 495-4403 Voice
301 495-4404 Fax
AIM: RandyAADB

Want to keep up with the latest news in the Deaf-Blind Community? Consider
subscribing to the monthly newsletter, "AADB Today" at http://aadb.org. It's
free and AADB membership is not required.

-----Original Message-----
From: = EMAIL ADDRESS REMOVED =
[mailto: = EMAIL ADDRESS REMOVED = ] On Behalf Of Patrick H. Lauke
Sent: Wednesday, January 14, 2009 1:13 PM
To: WebAIM Discussion List
Subject: Re: [WebAIM] Inline Images and ALT text

John Foliot wrote:

> Contrast: "Jared Smith, M.S., is the Associate Director for WebAIM. Jared
> Smith" vs. "Jared Smith, M.S., is the Associate Director for WebAIM.
> [Photo - Jared Smith]" Most screen readers, unless configured for maximum
> verbosity, tend to ignore/not voice the square brackets, and so would say
> aloud: "Jared Smith, M.S., is the Associate Director for WebAIM. Image -
> Photo - Jared Smith".

In this case, even better: "Jared Smith, M.S., is the Associate Director
for WebAIM." and a nice alt="" in the image?

P
--
Patrick H. Lauke

From: Jared Smith
Date: Wed, Jan 14 2009 12:55PM
Subject: Re: A larger discussion (was RE: Inline Images and ALT text)
← Previous message | Next message →

John Foliot wrote:
> The inclusion of a photo of Jared is done deliberately. It is not a
> "decorative" image per-se, it is provided to allow folks to know what Jared
> looks like, so that (for example) when you run into him at a conference you
> can say hi to him, or whatever. As such, it is an "image of value" as
> opposed to something that is purely decorative in nature.

I would agree that the image is not decorative. However, I would argue
that the content of the image ("Jared Smith") is already presented in
text. Adding the alternative text would be redundant. If the purpose
of the image was truly to provide a description of myself, I'd add alt
text akin to, "A strapping, handsome man..." :-)

> It is also
> entirely feasible that a non-sighted user might want a copy of that image
> (for a report, to add to their social network page, to share with an
> associate, etc.), yet by not indicating that the image is part of the page
> (and using alt=""), we/you are deliberately "hiding" that information "of
> value" from the non-sighted user.

And it is also entirely feasible (and I would argue more likely) that
the user does not care nor want to be informed of the presence of an
image just because it's an image. I think your logic is a bit of a
stretch. There's nothing that precludes a screen reader user from
finding images that don't have alt text. A screen reader could be set
to have them identified when reading. By following your
recommendation, you are now forcing the identification of the image
upon users that may not want that information. Most blind users I've
spoken to do not want images identified just because they are present.
They want content - and in this case, the entirety of the content is
already presented in text.

I admit that this is an area of accessibility where there will never
be 100% agreement. WebAIM will be compiling results from our screen
reader survey in the next few weeks and I think they will generally
show that these types of things make absolutely no difference to
screen reader users. But they are sure fun for accessibility folks to
debate ad nauseum.

Jared Smith
WebAIM

From: Dean Hamack
Date: Wed, Jan 14 2009 2:00PM
Subject: Re: A larger discussion (was RE: Inline Images and ALT text)
← Previous message | Next message →

On 1/14/09 11:54 AM, "Jared Smith" < = EMAIL ADDRESS REMOVED = > wrote:

> And it is also entirely feasible (and I would argue more likely) that
> the user does not care nor want to be informed of the presence of an
> image just because it's an image.

I couldn't agree more Jared. I think it's pretty ridiculous to force people
to listen to a bunch of unnecessary and redundant information.

From: John Foliot
Date: Wed, Jan 14 2009 2:15PM
Subject: Re: A larger discussion (was RE: Inline Images and ALT text)
← Previous message | Next message →

Jared Smith wrote:
>
> I would agree that the image is not decorative. However, I would argue
> that the content of the image ("Jared Smith") is already presented in
> text. Adding the alternative text would be redundant. If the purpose
> of the image was truly to provide a description of myself, I'd add alt
> text akin to, "A strapping, handsome man..." :-)

2 points: 1) that would be more appropriate as the longdesc text, 2) proof
of the subjectiveness of "appropriate" alt text... (<huge grin>)

> > It is also
> > entirely feasible that a non-sighted user might want a copy of that
image
> > (for a report, to add to their social network page, to share with an
> > associate, etc.), yet by not indicating that the image is part of the
page
> > (and using alt=""), we/you are deliberately "hiding" that information
"of
> > value" from the non-sighted user.
>
> And it is also entirely feasible (and I would argue more likely) that
> the user does not care nor want to be informed of the presence of an
> image just because it's an image.

Perhaps, but it is not *just* an image, it is an image that has some value
to it - it is being provided for a *reason*, else why insert it into the
page. It is directly related to, and in support of, the content it is
associated to.


> I think your logic is a bit of a
> stretch. There's nothing that precludes a screen reader user from
> finding images that don't have alt text.

...and web accessibility is more than just screen readers. C'mon Jared...
(<smile>)

> A screen reader could be set
> to have them identified when reading.

As an image, yes. However without actual text-value(s) associated to the
image, how does the user *know* that it is a photo of you, and not a
non-valuable decorative image? http://www.webaim.org/contact has 8
different biographical photos (out of 30 images present - 14 inline and 16
"background") Why make this difficult on the non-sighted user?

WCAG2.0:
Perceivable - Information and user interface components must be
presentable to users in ways they can perceive.
* This means that users must be able to perceive the information
being presented (it can't be invisible to all of their senses)

Without reconfiguring their AT, by default images with alt="" are not
identified to the end user - it is left out of the audio stream (or
Braille output).

> By following your
> recommendation, you are now forcing the identification of the image
> upon users that may not want that information.

You are forcing that picture on sighted users who may not want it
either... (not that it's a bad picture <grin>). I find this argument weak
- I have also heard from screen reader users that they want to be informed
of "significant" information, and that they want to decide for themselves
what is important or not, they prefer not to leave it to the "webmaster"
to decide... If a screenreader user "really" does not want to be informed
of inline images, they should configure *their* setup to behave this way.
(In W3C parlance, this is known as "author proposes, user disposes") And
by rights, if your photograph was non-essential to the content it should
be rendered via CSS as a background image: for the most part, alt="" is an
old-school hack that emerged back when we had no other way of solving this
issue.

> Most blind users I've
> spoken to do not want images identified just because they are present.

However in this instance, it's not just *any* image, again, in context, it
is an associated image of "value". It's not some swirly fancy bit of
eye-candy, it's your "official" bio photo - the same photo used in
numerous other web pages and related marketing pieces used at WebAIM. By
virtue of that, it has importance, and I believe you do a disservice to
the non-sighted by not specifically announcing its presence on the page.

> They want content - and in this case, the entirety of the content is
> already presented in text.

It is? I disagree. Your picture is an important part of the overall
content of the bio (else why include it?), yet for the non-sighted, you
have specifically not made its presence easily discernable... oh sure,
they can re-configure their AT to seek out the image here, but that is an
undue burden on them.

>
> I admit that this is an area of accessibility where there will never
> be 100% agreement.

Here we agree. (I also miss the days when we had larger, open discussions
such as this... web accessibility lists have become quite dry lately).
Hopefully the survey results will help us all (and I am very thankful that
WebAIM took this initiative for long overdue research).

> WebAIM will be compiling results from our screen
> reader survey in the next few weeks and I think they will generally
> show that these types of things make absolutely no difference to
> screen reader users.

Really? I suspect that it will not be so Black and White... again, when I
discuss this type of thing with screen reader users, most often I get a
50/50 split, but my pool of interviewees is far smaller than what your
survey is (hopefully) attracting, so time will tell. However, I think
that it is important that these types of "subjective" topics are also
discussed within experts communities, as from those types of discussions
emerge new thoughts and ideas, and that's always a good thing.


JF

From: Patrick H. Lauke
Date: Wed, Jan 14 2009 2:40PM
Subject: Re: A larger discussion (was RE: Inline Images and ALT text)
← Previous message | Next message →

John Foliot wrote:

>> I think your logic is a bit of a
>> stretch. There's nothing that precludes a screen reader user from
>> finding images that don't have alt text.
>
> ...and web accessibility is more than just screen readers. C'mon Jared...
> (<smile>)

But in this specific case, how are users with other
abilities/disabilities affected by an empty alt?

> However in this instance, it's not just *any* image, again, in context, it
> is an associated image of "value". It's not some swirly fancy bit of
> eye-candy, it's your "official" bio photo - the same photo used in
> numerous other web pages and related marketing pieces used at WebAIM. By
> virtue of that, it has importance, and I believe you do a disservice to
> the non-sighted by not specifically announcing its presence on the page.

What value do they draw from it, though? The case you cited ("a
non-sighted user might want a copy of that image (for a report, to add
to their social network page, to share with an associate, etc.)") seems
a bit of an edge case for me, but that's my personal opinion.

Happy to agree to disagree though :)

I'll throw in WAI-ARIA "labeledby" and "describedby" as an "alternative
to alt", if you will...
http://www.w3.org/TR/wai-aria/#labelledby
http://www.w3.org/TR/wai-aria/#describedby
(although yes, a static image on a static page may not be the intended
"Rich Internet Application" usage)

P
--
Patrick H. Lauke

From: John E. Brandt
Date: Wed, Jan 14 2009 2:50PM
Subject: Re: A larger discussion (was RE: Inline Images and ALT text)
← Previous message | Next message →

I ditto this...And I look forward to the research.

I've told this story before, so I will be brief...In looking at discussions
in this list over the years where folks who are screen reader users have
commented on the efficacy of "null alts" vs. alt descriptions, I concluded
long ago that there is no definitive answer to this question. The best I
could surmise was that for people who had some limited sight and/or had been
born with sight and developed vision problems later in life, images were
important. For others, they were not. But these were always opinions and
clearly not a scientific or authoritative analysis.

I look forward to something more than simple opinion to be offered, but I
suspect that many people will do what they want anyway, and perhaps that's
okay too.

Keeping warm in Maine...best to all!

~j

John E. Brandt
Web Design, Development, Consultation
Augusta, Maine USA
www.jebswebs.com
= EMAIL ADDRESS REMOVED =
207-622-7937 

-----Original Message-----
From: = EMAIL ADDRESS REMOVED =
[mailto: = EMAIL ADDRESS REMOVED = ] On Behalf Of Dean Hamack
Sent: Wednesday, January 14, 2009 3:43 PM
To: WebAIM Discussion List
Subject: Re: [WebAIM] A larger discussion (was RE: Inline Images and ALT
text)

On 1/14/09 11:54 AM, "Jared Smith" < = EMAIL ADDRESS REMOVED = > wrote:

> And it is also entirely feasible (and I would argue more likely) that
> the user does not care nor want to be informed of the presence of an
> image just because it's an image.

I couldn't agree more Jared. I think it's pretty ridiculous to force people
to listen to a bunch of unnecessary and redundant information.

From: John Foliot
Date: Wed, Jan 14 2009 5:45PM
Subject: Re: A larger discussion (was RE: Inline Images and ALT text)
← Previous message | Next message →

Patrick H. Lauke wrote:
>
> But in this specific case, how are users with other
> abilities/disabilities affected by an empty alt?

OK, I'll bite just to keep discussion going (for those lurkers out there).
And let me state that this is also coming from the "Universal Design"
perspective rather than the "accommodation" perspective (either point of
view being germane, neither being wrong)

Scenario: dial-up user in Gopher Gulch USA has images disabled for improved
page load speed. In this case, most browsers will instead show the alt text
associated to the image, and if I, as that Gopher Gulch user decide that I
want to see what Jared looks like (but am less inclined to see what John
Whiting looks like), I can choose to download simply the single image.
Without the presence of alternative text, all I am currently getting is some
additional white space.
(A stretch? Perhaps, but not too much of one. And I also refer once again
to the W3C notion of "author proposes, *user* disposes" as being a best
practices position). As educators/advocates/community leaders, I also
believe that it is important that we walk the walk as well as talking the
talk). As Jared noted, there will likely never be 100% agreement here, as
at the end of the day this is something of a subjective decision. But for
web content, end users can always remove bits of content, but they can never
add bits not supplied.

>
> What value do they draw from it, though? The case you cited ("a
> non-sighted user might want a copy of that image (for a report, to add
> to their social network page, to share with an associate, etc.)") seems
> a bit of an edge case for me, but that's my personal opinion.

<important>
The value (or lack of value) they get from knowing specifically that there
is an image of Jared Smith on that page is equal to the value that any other
sighted user may get - it ranges from zero to 100%; to me what is most
important is that I am not treating the non-sighted user any differently
than the sighted user by deciding to show the image to the sighted, but hide
it from the non-sighted using alt="".
</important>

Why should *I* as webmaster decide what is or isn't important to any given
user? As I said earlier, if it is truly a "throw-away" image, it should be
added to the page as a CSS background. If it is important enough for the
webmaster to include on the page as an inline image, it should be important
enough to inform all users that the image is there (from the get-go).
Having them re-adjust their UA/AT to "discover" images with alt="" is simply
not right (IMHO)

>
> Happy to agree to disagree though :)

Well, it's not so much disagree as talking through the idea fully. I
certainly understand where you and Jared are coming from, and since both of
you are certainly acknowledged web accessibility "experts" (and without any
false modesty, I kinda think I know of what I speak too), I think that it is
important that we air these ideas in public forums (where the majority
simply lurk and learn). I am advocating a position, putting forth my
justifications, and letting readers think and decide for themselves. So are
you, so this is good.

>
> I'll throw in WAI-ARIA "labeledby" and "describedby" as an "alternative
> to alt", if you will...
> http://www.w3.org/TR/wai-aria/#labelledby
> http://www.w3.org/TR/wai-aria/#describedby
> (although yes, a static image on a static page may not be the intended
> "Rich Internet Application" usage)

You had to remind me of HTML5 huh <grin>. You're right in principle,
although is the browser/AT support for this really Primetime yet? I'm
thinking that it will be a few years still before we can reliably rely on
these ARIA attributes consistently although that should not stop developers
from implementing ARIA attributes now.

JF

From: Léonie Watson
Date: Thu, Jan 15 2009 2:00AM
Subject: Re: A larger discussion (was RE: Inline Images and ALT text)
← Previous message | Next message →

"And it is also entirely feasible (and I would argue more likely) that the user does not care nor want to be informed of the presence of an image just because it's an image. I think your logic is a bit of a stretch."

As has been said in this thread elsewhere, we're unlikely to find full agreement on this. People who lost their sight during their lifetime are likely to respond differently to alt descriptions, than people who were born without sight.

As someone who falls into the former category, and still visualises a great deal, I find that alt descriptions often act as cues to my imagination, as much as they provide an accurate and specific description of something. Take an image with an alt text of "Jared Smith", on a profile page...

The alt text doesn't tell me much about the way Jared looks, but it does give me a cue to visualise a typical profile page in my head. This doesn't deal with specifics, but it does add to my overall sense of how the page might look. It might also give me a cue to consider what Jared actually does look like (tall and strapping I've heard ;). This too can add to the experience of visiting a page, although not perhaps in the conventional way intended by alt descriptions.


"I admit that this is an area of accessibility where there will never be 100% agreement. WebAIM will be compiling results from our screen reader survey in the next few weeks and I think they will generally show that these types of things make absolutely no difference to screen reader users. But they are sure fun for accessibility folks to debate ad nauseum."


Here I suspect you're right. I completed the survey, but don't recall whether people were asked about the time when they lost their sight. I think this would be a critical factor in trying to find a common ground with the issue of alt texts. Very much looking forward to the results though.



Regards,
Léonie.

--
Nomensa - humanising technology

Léonie Watson | Director of Accessibility
t. +44 (0)117 929 7333 |

-----Original Message-----
From: = EMAIL ADDRESS REMOVED = [mailto: = EMAIL ADDRESS REMOVED = ] On Behalf Of Jared Smith
Sent: 14 January 2009 19:54
To: WebAIM Discussion List
Subject: Re: [WebAIM] A larger discussion (was RE: Inline Images and ALT text)

John Foliot wrote:
> The inclusion of a photo of Jared is done deliberately. It is not a
> "decorative" image per-se, it is provided to allow folks to know what
> Jared looks like, so that (for example) when you run into him at a
> conference you can say hi to him, or whatever. As such, it is an
> "image of value" as opposed to something that is purely decorative in nature.

I would agree that the image is not decorative. However, I would argue that the content of the image ("Jared Smith") is already presented in text. Adding the alternative text would be redundant. If the purpose of the image was truly to provide a description of myself, I'd add alt text akin to, "A strapping, handsome man..." :-)

> It is also
> entirely feasible that a non-sighted user might want a copy of that
> image (for a report, to add to their social network page, to share
> with an associate, etc.), yet by not indicating that the image is part
> of the page (and using alt=""), we/you are deliberately "hiding" that
> information "of value" from the non-sighted user.

And it is also entirely feasible (and I would argue more likely) that the user does not care nor want to be informed of the presence of an image just because it's an image. I think your logic is a bit of a stretch. There's nothing that precludes a screen reader user from finding images that don't have alt text. A screen reader could be set to have them identified when reading. By following your recommendation, you are now forcing the identification of the image upon users that may not want that information. Most blind users I've spoken to do not want images identified just because they are present.
They want content - and in this case, the entirety of the content is already presented in text.

I admit that this is an area of accessibility where there will never be 100% agreement. WebAIM will be compiling results from our screen reader survey in the next few weeks and I think they will generally show that these types of things make absolutely no difference to screen reader users. But they are sure fun for accessibility folks to debate ad nauseum.

Jared Smith
WebAIM

From: Léonie Watson
Date: Thu, Jan 15 2009 2:10AM
Subject: Re: A larger discussion (was RE: Inline Images and ALT text)
← Previous message | Next message →

"I couldn't agree more Jared. I think it's pretty ridiculous to force people to listen to a bunch of unnecessary and redundant information."

Why would you be forcing someone to listen to it?



Regards,
Léonie.

--
Nomensa - humanising technology

Léonie Watson | Director of Accessibility
t. +44 (0)117 929 7333 |

-----Original Message-----
From: = EMAIL ADDRESS REMOVED = [mailto: = EMAIL ADDRESS REMOVED = ] On Behalf Of Dean Hamack
Sent: 14 January 2009 20:43
To: WebAIM Discussion List
Subject: Re: [WebAIM] A larger discussion (was RE: Inline Images and ALT text)

On 1/14/09 11:54 AM, "Jared Smith" < = EMAIL ADDRESS REMOVED = > wrote:

> And it is also entirely feasible (and I would argue more likely) that
> the user does not care nor want to be informed of the presence of an
> image just because it's an image.

I couldn't agree more Jared. I think it's pretty ridiculous to force people to listen to a bunch of unnecessary and redundant information.

From: M Akram Danish
Date: Thu, Jan 15 2009 2:35AM
Subject: Re: Inline Images and ALT text
← Previous message | Next message →

> In this case, even better: "Jared Smith, M.S., is the Associate Director for WebAIM."
> and a nice alt="" in the image?

from readability point it is good, but it may not let the visitor know there is a photo of Jared Smith. Even thought he she not see the photo but aware the existence and can tell that the photo exist there and if needed any sight person can check.

but if that is just a decoration piece then sure your example seems good. just last day I talked with my visually impaired student and she said she prefer null alt for decoration images. When I asked won't it make you curious what this image is for she said her personal preference is alt = ""

Akram




> Contrast:  "Jared Smith, M.S., is the Associate Director for WebAIM. Jared
> Smith" vs. "Jared Smith, M.S., is the Associate Director for WebAIM.
> [Photo - Jared Smith]"  Most screen readers, unless configured for maximum
> verbosity, tend to ignore/not voice the square brackets, and so would say
> aloud: "Jared Smith, M.S., is the Associate Director for WebAIM. Image -
> Photo - Jared Smith".

In this case, even better: "Jared Smith, M.S., is the Associate Director for WebAIM." and a nice alt="" in the image?

P
-- Patrick H. Lauke

From: Oliver Boermans
Date: Thu, Jan 15 2009 3:30AM
Subject: Re: A larger discussion (was RE: Inline Images and ALT text)
← Previous message | Next message →

Thanks for the interesting discussion people!

As one of those learning lurkers [thanks John], I'd like to put
forward a design/communication principal which may have some bearing
on ensuring content is accessible and interesting.

With an idea to express, the most effective combination of text and
image is one where the visual and verbal elements work together. An
images with a caption that seems to restate the apparent content of
the image is boring alongside one that has a caption that tells us
something more. The viewers is involved by connecting the two. Having
a little gap is what keeps us engaged.

When the image is not visible it is important to consider what is
missing from the message as a result. I believe it is the role of the
alt text to at least attempt to adopt the role of the image in the
communication.

With a different hat on, it is easy for me to imagine scenarios where
having some such redundancy within the alt text of an image would be
beneficial. Placing an image on a public webpage is also placing an
image on the web. A web increasingly read by machines at some point
before we are presented with it. Simplest example is Google images
which works best when the information within the alt text clearly
labels the content of the image.

Mark me as somewhere in the middle – it's a balancing act.

--
Oliver Boermans
Communication designer
http://www.ollicle.com
--

From: Dan Conley
Date: Thu, Jan 15 2009 7:40AM
Subject: Re: A larger discussion (was RE: Inline Images and ALT text)
← Previous message | Next message →

This right here is why I subscribe to the list!

Almost all of our images fall into two categories: decorative and
complex. I use an empty alt for the (few) decorative images we have, and
the rest use long descriptions linked to in the text, so essentially all
of our images have empty alts. I see what people mean about the image
disappearing, though, and so I think I'm going to go through and
replace, for example, the empty alt text in Figure 1 of an encyclopedia
article with 'Figure 1.' Immediately after the image there's a link to
the description, if they want to know what it's saying, and if a text
user, etc, wants the image they know where to get it.

Right?

(that's the rub with accessibility: like most things in life, there's no
one completely right answer)

Dan Conley
Information Specialist
Center for International Rehabilitation Research Information and
Exchange (CIRRIE)
University at Buffalo, Health Sciences Library B6
Phone: (716) 829-3900 x145
= EMAIL ADDRESS REMOVED =
http://cirrie.buffalo.edu

Oliver Boermans wrote:
> Thanks for the interesting discussion people!
>
> As one of those learning lurkers [thanks John], I'd like to put
> forward a design/communication principal which may have some bearing
> on ensuring content is accessible and interesting.
>
> With an idea to express, the most effective combination of text and
> image is one where the visual and verbal elements work together. An
> images with a caption that seems to restate the apparent content of
> the image is boring alongside one that has a caption that tells us
> something more. The viewers is involved by connecting the two. Having
> a little gap is what keeps us engaged.
>
> When the image is not visible it is important to consider what is
> missing from the message as a result. I believe it is the role of the
> alt text to at least attempt to adopt the role of the image in the
> communication.
>
> With a different hat on, it is easy for me to imagine scenarios where
> having some such redundancy within the alt text of an image would be
> beneficial. Placing an image on a public webpage is also placing an
> image on the web. A web increasingly read by machines at some point
> before we are presented with it. Simplest example is Google images
> which works best when the information within the alt text clearly
> labels the content of the image.
>
> Mark me as somewhere in the middle – it's a balancing act.
>
> --
> Oliver Boermans
> Communication designer
> http://www.ollicle.com
> --
>

From: J. B-Vincent
Date: Fri, Jan 16 2009 9:55AM
Subject: Re: A larger discussion (was RE: Inline Images and ALT text)
← Previous message | Next message →

A modest proposal...

What if there were a way for users to specify how they wanted to perceive ALT attributes, possibly at first through screen reader settings (similar to how screen reader users can set levels for how much punctuation they want to hear) and eventually through the browser. Levels might include:

* None--the user perceives no ALT attributes; good for screen reader users who want to move through pages as fast as possible.

* Some--the user perceives non-null ALT attributes; good for anyone who wants a sense of graphics location and content within the page.

* All--the user perceives all ALT attributes; good for sighted users with low bandwidth who might deal with a lot of pages that have info-bearing pictures such as charts.

The latter two levels might assign an automatic null value for any graphics that omit the ALT attribute.

At an even higher level, users might be able to mark preferences for specific sites. For example, screen reader users could set a news site to "none" but an online learning site to "all."

There would still need to be website designer awareness on how to write good ALT attributes, but this capability could give _all_ users more control over page presentation regardless of designers' awareness or choices.

--Jane

*****

Jane Vincent
Accessibility/Usability Manager
Center for Accessible Technology
Berkeley, CA


--- On Thu, 1/15/09, Oliver Boermans < = EMAIL ADDRESS REMOVED = > wrote:
From: Oliver Boermans < = EMAIL ADDRESS REMOVED = >
Subject: Re: [WebAIM] A larger discussion (was RE: Inline Images and ALT text)
To: "WebAIM Discussion List" < = EMAIL ADDRESS REMOVED = >
Date: Thursday, January 15, 2009, 2:25 AM

Thanks for the interesting discussion people!

As one of those learning lurkers [thanks John], I'd like to put
forward a design/communication principal which may have some bearing
on ensuring content is accessible and interesting.

With an idea to express, the most effective combination of text and
image is one where the visual and verbal elements work together. An
images with a caption that seems to restate the apparent content of
the image is boring alongside one that has a caption that tells us
something more. The viewers is involved by connecting the two. Having
a little gap is what keeps us engaged.

When the image is not visible it is important to consider what is
missing from the message as a result. I believe it is the role of the
alt text to at least attempt to adopt the role of the image in the
communication.

With a different hat on, it is easy for me to imagine scenarios where
having some such redundancy within the alt text of an image would be
beneficial. Placing an image on a public webpage is also placing an
image on the web. A web increasingly read by machines at some point
before we are presented with it. Simplest example is Google images
which works best when the information within the alt text clearly
labels the content of the image.

Mark me as somewhere in the middle – it's a balancing act.

--
Oliver Boermans
Communication designer
http://www.ollicle.com
--

From: John Foliot
Date: Fri, Jan 16 2009 10:50AM
Subject: Re: A larger discussion (was RE: Inline Images and ALT text)
← Previous message | Next message →

J. B-Vincent wrote:
>
> What if there were a way for users to specify how they wanted to
> perceive ALT attributes, possibly at first through screen reader
> settings (similar to how screen reader users can set levels for how
> much punctuation they want to hear) and eventually through the browser.

Jane,

While this proposal certainly has merit, it also re-enforces the fact that
the decision process should (must?) remain with the end user. It further
supports the W3C concept of "author proposes, user disposes". As content
authors then, it also cements the idea that we should be providing content
for the alt value (and not "") whenever possible... let the user decide to
accept or ignore this data, don't presume that they may not want to have
it, based upon *your* perception of user needs.

>
> The latter two levels might assign an automatic null value for any
> graphics that omit the ALT attribute.

The debates surround error recovery for images that lack any alt value
could fill volumes (hint: search out the HTML5 archives), and is far
beyond what this thread could handle. I would suggest that this is not
necessarily the best solution, but if you think the pros and cons of the
current thread are wide-ranging, you ain't seen nothing 'till you delve
into that discussion.

>
> At an even higher level, users might be able to mark preferences for
> specific sites. For example, screen reader users could set a news site
> to "none" but an online learning site to "all."

Indeed, this type of granularity would be extremely useful. The WAI ARIA
spec has a capacity to determine a level of "politeness" (for WAI ARIA
Live Regions http://tinyurl.com/8j32zm) which has a similar kind of
switching mechanism, but I am not sure if there has been any work done
along the lines you propose.

>
> There would still need to be website designer awareness on how to write
> good ALT attributes, but this capability could give _all_ users more
> control over page presentation regardless of designers' awareness or
> choices.

<opinion>
Hear hear! While 3 web accessibility experts in the same room will likely
have 3 differing but correct opinions of what is "appropriate (or good)
alt text" for any given image[1], I would suggest that a null value is
usually never appropriate (or at least, that's the position I've had in
this thread from the start). If the image does not really require alt
text, it should be in your CSS, and not inline.
</opinion>


JF
============================
John Foliot
Program Manager
Stanford Online Accessibility Program
http://soap.stanford.edu
Stanford University
Tel: 650-862-4603

Soap Is a program directed by the
Vice Provost for Student Affairs
============================

[1] Good geeky fun when web accessibility folk get together - grab a
selection of images and have everyone voice their idea of "good or
appropriate" alt text. Losers pay the round... (LOL)

From: Randall Pope
Date: Fri, Jan 16 2009 11:15AM
Subject: Re: A larger discussion (was RE: Inline Images and ALTtext)
← Previous message | Next message →

Hi John,

<opinion>
Hear hear! While 3 web accessibility experts in the same room will likely
have 3 differing but correct opinions of what is "appropriate (or good)
alt text" for any given image[1], I would suggest that a null value is
usually never appropriate (or at least, that's the position I've had in
this thread from the start). If the image does not really require alt
text, it should be in your CSS, and not inline.
</opinion>

Hear hear! I seconded this approach. To be honest, I'm at lost why many
webmasters do not place the images, that don't required alt text, in the
CSS. Could someone explain why this approach is not being use often? I
have been asked by screen readers numerous times to check a website's image
only to find it was being use for decorated purpose.

With Warm Regards,
Randall "Randy" Pope
American Association of the Deaf-Blind
Website: http://www.aadb.org

301 495-4402 VP/TTY
301 495-4403 Voice
301 495-4404 Fax
AIM: RandyAADB

Want to keep up with the latest news in the Deaf-Blind Community? Consider
subscribing to the monthly newsletter, "AADB Today" at http://aadb.org. It's
free and AADB membership is not required.


<opinion>
Hear hear! While 3 web accessibility experts in the same room will likely
have 3 differing but correct opinions of what is "appropriate (or good)
alt text" for any given image[1], I would suggest that a null value is
usually never appropriate (or at least, that's the position I've had in
this thread from the start). If the image does not really require alt
text, it should be in your CSS, and not inline.
</opinion>


JF
============================
John Foliot
Program Manager
Stanford Online Accessibility Program
http://soap.stanford.edu
Stanford University
Tel: 650-862-4603

Soap Is a program directed by the
Vice Provost for Student Affairs
============================

[1] Good geeky fun when web accessibility folk get together - grab a
selection of images and have everyone voice their idea of "good or
appropriate" alt text. Losers pay the round... (LOL)

From: John Foliot
Date: Fri, Jan 16 2009 11:30AM
Subject: Re: A larger discussion (was RE: Inline Images and ALTtext)
← Previous message | Next message →

Randall Pope wrote:
>
> > If the image does not really require alt
> > text, it should be in your CSS, and not inline.
>
> Hear hear! I seconded this approach. To be honest, I'm at lost why many
> webmasters do not place the images, that don't required alt text, in
> the
> CSS. Could someone explain why this approach is not being use often?

LOL... well, this is purely opinion, but I suspect the answer is
"ignorance" (and I don't mean this maliciously).

Until recently (4-6 years?), good CSS support was hard to rely on. Legacy
content, legacy development skills (they learned Dreamweaver in 2001, and
have been using it ever since in exactly the same fashion), WYSIWYG
editors (drag the image to the place on the page you want it), and more...

The sad reality is that most web authors simply are unaware of evolving
best practices, and when you spend less than 10 hours a week updating your
website, there is just no reason to do things differently than what you
originally learned. Hopefully over time this will improve, and it takes
people like those subscribed to this list to further teach your peers,
friends, and work associates how to improve their skills to not only
benefit the disabled community, but themselves as well.

> I
> have been asked by screen readers numerous times to check a website's
> image
> only to find it was being use for decorated purpose.

Yet one more reason to not use alt=""...

JF

From: Jared Smith
Date: Fri, Jan 16 2009 11:55AM
Subject: Re: A larger discussion (was RE: Inline Images and ALTtext)
← Previous message | Next message →

On Fri, Jan 16, 2009 at 11:10 AM, Randall Pope < = EMAIL ADDRESS REMOVED = > wrote:
> To be honest, I'm at lost why many
> webmasters do not place the images, that don't required alt text, in the
> CSS. Could someone explain why this approach is not being use often?

A lot (most?) of these images are floated beside the content with
which they are associated. It's impossible in CSS to float content
around a background image without inserting an extra element (maybe an
empty span or div) into the content. And if you're going to add a
worthless element into your content, why not just add the image
itself? The truth is, there just isn't nearly enough flexibility in
CSS for positioning background images and content in a way that works.

Take the page in question (http://webaim.org/contact/) - there is no
way to put the photos in CSS without inserting at least one
additional, empty element (with it's own CSS definitions) to take up
their space.

And as far as functional accessibility goes, there really isn't a heap
of difference between an image with alt="" and a CSS background image.
I'm not arguing that the CSS isn't better - it is, but the argument
that it is necessary or even better for accessibility isn't a very
strong one.

Jared Smith
WebAIM

From: John Foliot
Date: Fri, Jan 16 2009 2:05PM
Subject: Re: A larger discussion (was RE: Inline Images and ALTtext)
← Previous message | Next message →

Jared Smith wrote:
> And if you're going to add a
> worthless element into your content, why not just add the image
> itself?
>
> Take the page in question (http://webaim.org/contact/) - there is no
> way to put the photos in CSS without inserting at least one
> additional, empty element (with it's own CSS definitions) to take up
> their space.

Jared,

If the biographical photos of you and the staff are "worthless", then why
are you adding them to the page? There is a value-add proposition for
including those photos, and with alt="" you are denying non-sighted users
the ability to access or discard that value (unless they jump through
hoops and re-configure their AT to find images with alt="" - at which
point some users then ask people like Randall to tell them what the image
is). This is the crux of the debate.

The photos in question are *NOT* non-essential elements on the page - a
conscious decision was taken to arrange for the photos to be supplied
(likely even engaging a photographer and all staff assembling on a given
day to have pictures taken - they don't strike me as being candid shots),
they were then specifically added to the bio page in association with the
textual content, and they form part of the over-all "message" being
delivered to the majority of users. To suggest that they might add audio
clutter to the content for non-sighted users is a weak argument, and
presumes that you know what all users accessing that page
need/want/require. Are the images "visual clutter"?

When an image reaches the status of requiring in-line placement (so that
you can float it, etc.), then that image has reached the status of having
some "importance" (how "important" remains open for discussion, but surely
higher than background images - which sit in the CSS). I maintain then
that if it is thus important, there should be a value to the @alt
attribute beyond null. If we were to push the envelope just a tad, I
would further suggest that images of highest level importance should
likely also have a longdesc value as well (for the strapping, handsome men
out there).

>
> And as far as functional accessibility goes, there really isn't a heap
> of difference between an image with alt="" and a CSS background image.
> I'm not arguing that the CSS isn't better - it is, but the argument
> that it is necessary or even better for accessibility isn't a very
> strong one.

Agreement here, although methinks that using CSS is a better development
practice. I'd like to see alt="" disappear, and the sooner we start doing
just that, the sooner we will reach that point.

JF
============================
John Foliot
Program Manager
Stanford Online Accessibility Program
http://soap.stanford.edu
Stanford University
Tel: 650-862-4603

Soap Is a program directed by the
Vice Provost for Student Affairs
============================

From: S.R. Emerson, Accrete Web Solutions
Date: Fri, Jan 16 2009 2:30PM
Subject: Re: A larger discussion (was RE: Inline Images and ALTtext)
← Previous message | Next message →

>> Hear hear! I seconded this approach. To be honest, I'm at lost why many
>> webmasters do not place the images, that don't required alt text, in
>> the
>> CSS. Could someone explain why this approach is not being use often?

Maybe because the images will not be printed if the page is printed? By
default, the browser does not print backgrounds.

e.g. If you insert the logo via CSS, it won't be printed when the user
prints the page
A logo would be a important image to have on the printed version of the page
but, if the header is coded first in the background coding c/w the alt
attribute filled in, that's not good for SEO.

If you know how to code the page with the content first and manipulate the
header via CSS to visually appear at the top of the page, then you satisfy
solving a SEO problem and the image can then have an alt attribute that is
not empty.

S. Emerson
Accrete Web Solutions
http://www.accretewebsolutions.ca

From: John Foliot
Date: Fri, Jan 16 2009 2:35PM
Subject: Re: A larger discussion (was RE: Inline Images and ALTtext)
← Previous message | Next message →

S.R. Emerson, Accrete Web Solutions wrote:
>
>
> Maybe because the images will not be printed if the page is printed?
> By
> default, the browser does not print backgrounds.

Ergo, the images are important. One more reason why they should have an
alt value greater than null: if they are important enough to "print" (a
visual concept), then they are likely important enough to "hear" as
well...


JF

From: Jared Smith
Date: Fri, Jan 16 2009 3:10PM
Subject: Re: A larger discussion (was RE: Inline Images and ALTtext)
← Previous message | Next message →

On Fri, Jan 16, 2009 at 2:03 PM, John Foliot < = EMAIL ADDRESS REMOVED = > wrote:

> If the biographical photos of you and the staff are "worthless", then why
> are you adding them to the page? There is a value-add proposition for
> including those photos, and with alt="" you are denying non-sighted users
> the ability to access or discard that value.

Once again, I am not arguing that the images are worthless,
non-essential, or don't convey something - content even. What I am
saying is that the content of my photograph is nothing more than
"Jared Smith" and that content is already fully conveyed in text
adjacent to the image. So why force this redundancy upon the user? If
I believed the content were truly something more than "Jared Smith" (I
can't imagine what - unless I were to *describe* the image - something
typically worse than useless), I would have added it in alt in the
first place.

If you remove the image altogether is any important *content* lost? I
would say no. You would probably disagree.

Yes, you may find screen reader (or other) users that would want to
hear or see "Jared Smith. Photo of Jared Smith.", but I think this is
the exception rather than the norm. In short, you have to weigh the
possible advantage of identifying the presence of the image with the
possible disadvantage of redundancy and over-burdening the user with
information. For me, the disadvantages outweigh the potential
advantage.

Is the user experience significantly different if you add brief alt
text, add alt="", or put it in a CSS background? No! It just doesn't
matter that much. But that doesn't mean it's not fun to debate such
things.

Jared

P.S. I've never referred to myself in the 3rd person so much in my life!

From: John Foliot
Date: Fri, Jan 16 2009 3:45PM
Subject: Re: A larger discussion (was RE: Inline Images and ALT text)
← Previous message | Next message →

Jared Smith wrote:
>
> Yes, you may find screen reader (or other) users that would want to
> hear or see "Jared Smith. Photo of Jared Smith.", but I think this is
> the exception rather than the norm.

In the context of this debate/discussion, two respondents have opined that
having the value of the alt text in this particular situation have a value
greater than _null_ is the preferred option. Léonie Watson (a
self-identified screen reader user) noted that alt values helped her
"visualize" the page, and would appreciate knowing that "Photo - Jared
Smith" existed. Randall Pope provided anecdotal evidence that he has "...
been asked by screen readers numerous times to check a website's image
only to find it was being use for decorated purpose..." (thus alt=""
introduces ambiguity) and conversely none have suggested that in this
example not being signaled that your photo exists is the preference.

So while we anxiously await the results of your poll (and I am *not* being
facetious, I really welcome this kind of quantifiable data - we need more
of it), our extremely small straw poll suggests that perhaps you may be
mistaken here.

> In short, you have to weigh the
> possible advantage of identifying the presence of the image with the
> possible disadvantage of redundancy and over-burdening the user with
> information. For me, the disadvantages outweigh the potential
> advantage.

I suppose then my friend that at the end of the day we will continue to
agree to disagree. I believe that your judgment regarding the
disadvantage of redundancy vs. signaling the presence of the image is
over-rated, but at the end of the day it really comes down to individual
users - I'm sure if we searched hard enough we could both find a football
team's worth of users that lined up on either side of the discussion. If
nothing it serves to prove that real web accessibility is often
subjective, and that serious practitioners need to truly examine specific
details, and use-cases. Often there really is no right or wrong, but only
shades of gray.

Cheers!

JF

From: Chris Hoffman
Date: Sat, Jan 17 2009 1:30PM
Subject: Re: A larger discussion (was RE: Inline Images and ALT text)
← Previous message | No next message

What about using non-null alt attributes with WAI-ARIA's
role="presentation"? In theory that would mark images as
"nonessential", but would also provide descriptions for anyone who
wanted them. In practice, I have no idea. Maybe someone else can
report on how major screen readers deal with presentation roles on
images?