WebAIM - Web Accessibility In Mind

E-mail List Archives

Thread: FW: alt text size

for

Number of posts in this thread: 9 (In chronological order)

From: John Foliot
Date: Mon, Aug 21 2006 12:30PM
Subject: FW: alt text size
No previous message | Next message →

zara wrote:
> Hi,

And hi to you. Jukka gave a fairly full and concise explanation, so I won't
re-tread the same water; however I suspect that one of the things that is
being considered is the "tool-tip" feature of Internet Explorer attached to
alt text. While this should not be an "expected" behaviour (as it does not
render the same in other browers) many developers use it to include
additional information attached to their image - they think they are doing
the right thing when in fact they are making it worse. (For what it's worth,
a quick check shows that IE will support a fairly long string of characters
in the "tool tip" - more than I would like to see. Opera supports a long
string of TITLE values as a "tool tip", whereas my copy of Firefox cut off
the TITLE text at about 65 characters +/-. NEITHER OF THESE METHODS SHOULD
BE USED TO EXCLUSIVELY CONVEY INFORMATION!)

So, rather than looking a specific numbers of either characters or words,
perhaps a completely different angle of attack should be considered. Alt
text should be kept under approximately 10 words, AND USED TO DESCRIBE WHAT
THE IMAGE IS, rather than what it means. If the image is sufficiently
complex that it requires further explanation then use the LONGDESC
attribute, or better yet, ensure that the corresponding text is visible on
the same page as the complex image (not always easy, I know).

Related to this topic - I have taken to using a particular construct when
declaring my alt texts: alt="[Photo - Catherine Roy]".

1) By using the square brackets, it keeps the alt text separated in text
display browsers (it's a visual thing), especially since sometimes the alt
text alone may seem or appear redundantly redundant:

Catherine Roy Catherine Roy is a wonderful...
vs.
[Photo - Catherine Roy] Catherine Roy is a wonderful...

Note: some screen readers (set to maximum verbosity/read all punctuation)
will announce the presence of the square brackets. It may be a bit much for
them, but by experience very few daily screen reader users have their setups
set to maximum verbosity most of the time. To my mind the pros outweigh the
cons.

2) I also specify the fact that it is a Photo; I also use Image and Icon
as descriptors - my thoughts are that they are consistent, concise and
descriptive: a general "what" followed by a concise explanation:

[Image - W3Q Logo]
[Icon - Adobe Acrobat]

Perhaps hard to mandate into a Standards document, but a consideration for
"best practices"?

Thoughts? Comments?

JF
---
John Foliot
Academic Technology Specialist - Online Accessibility
Stanford University
560 Escondido Mall
Meyer Library 181
Stanford, CA 94305-3093






From: Patrick H. Lauke
Date: Mon, Aug 21 2006 1:30PM
Subject: Re: FW: alt text size
← Previous message | Next message →

John Foliot wrote:

> So, rather than looking a specific numbers of either characters or words,
> perhaps a completely different angle of attack should be considered. Alt
> text should be kept under approximately 10 words, AND USED TO DESCRIBE WHAT
> THE IMAGE IS, rather than what it means.

I'd add "and/or what its function is, if it is used as something like a
button or link".

> Related to this topic - I have taken to using a particular construct when
> declaring my alt texts: alt="[Photo - Catherine Roy]".
>
> 1) By using the square brackets, it keeps the alt text separated in text
> display browsers (it's a visual thing),

To be hardline, it should be the browser's responsibility to show what
is an image alt and what is "normal" text.

P
--
Patrick H. Lauke
___________
re

From: John Foliot
Date: Mon, Aug 21 2006 1:40PM
Subject: RE: FW: alt text size
← Previous message | Next message →

Patrick H. Lauke wrote:
>>
>> 1) By using the square brackets, it keeps the alt text separated in
>> text display browsers (it's a visual thing),
>
> To be hardline, it should be the browser's responsibility to show what
> is an image alt and what is "normal" text.
>

...and you wouldn't get any argument from me Patrick. However, in the
current world... (and besides, I cannot see a downside to my "technique",
can you?)

Cheers!

JF







From: Jared Smith
Date: Mon, Aug 21 2006 1:50PM
Subject: Re: FW: alt text size
← Previous message | Next message →

John Foliot wrote:
> 2) I also specify the fact that it is a Photo; I also use Image and Icon
> as descriptors - my thoughts are that they are consistent, concise and
> descriptive: a general "what" followed by a concise explanation:
>
> [Image - W3Q Logo]
> [Icon - Adobe Acrobat]
>
> Perhaps hard to mandate into a Standards document, but a consideration for
> "best practices"?

Interesting approach with the square brackets. I've found that JAWS reads
"left bracket" and "right bracket" in the default configuration. While the
motivation for this is visual separation thing, I would think that this
should best be done by the user agent, not by the introduction of
extraneous characters to the content. However, some text-only browsers
don't provide this separation. My own *opinion* would be that the overhead
of the brackets and corruption (perhaps this is not the best word) of the
content provides a negative impact that outweighs the benefit the brackets
might give to those using text-only browsers (a very small population - I
would guess much smaller than the screen reader population).

While the inclusion of "photo", "painting", or "icon" might be appropriate
in cases where it is important for the user to know the nature of the
image (for instance, a photo as compared to a painting), I think most
people would argue against identifying images as images within alternative
text. And I can't think of any case where the word "image" would be
appropriate within alternative text. The purpose of alternative text is to
provide an alternative to the content and/or function of the image, not to
identify the presence of an image.

As far as length, I think identifying a hard maximum is dangerous. My own
definition for alternative text length would be, "as succinctly as is
appropriate and no more than a short sentence or two." Anything more
concrete than that will likely result in insufficient or inappropriate alt
text. More important would be instructing what alt text really is and how
it is used.

I do find it interesting that there is still such a divergent opinion of
best practices for alternative text, the first principle of web
accessibility. I'm close to completing an article that will present my own
opinion of some best practices.

Jared Smith
WebAIM.org





From: Patrick H. Lauke
Date: Mon, Aug 21 2006 2:10PM
Subject: Re: FW: alt text size
← Previous message | Next message →

Jared Smith wrote:

> Interesting approach with the square brackets.

Wan't this something Joe Clark proposed (or at least documented) in his
book?

> the overhead of the brackets and corruption (perhaps this is not the
> best word) of the content

Yup, my thoughts exactly. I usually favour "sullying my content" ;-)

> most people would argue against identifying images as images
> within alternative text. And I can't think of any case where the word
> "image" would be appropriate within alternative text. The purpose of
> alternative text is to provide an alternative to the content and/or
> function of the image, not to identify the presence of an image.

Definitely. We're strangely in sync on this one, Jared.

> As far as length, I think identifying a hard maximum is dangerous. My
> own definition for alternative text length would be, "as succinctly as
> is appropriate and no more than a short sentence or two." Anything more
> concrete than that will likely result in insufficient or inappropriate
> alt text. More important would be instructing what alt text really is
> and how it is used.

Setting an arbitrary length in characters or words would definitely be
counter productive. It tries to pin hard and fast certainty where there
really needs to be case by case consideration.

P
--
Patrick H. Lauke
___________
re

From: John Foliot
Date: Mon, Aug 21 2006 2:20PM
Subject: RE: FW: alt text size
← Previous message | Next message →

Jared Smith wrote:
>
> Interesting approach with the square brackets. I've found that JAWS
> reads "left bracket" and "right bracket" in the default
> configuration. While the
> motivation for this is visual separation thing, I would think that
> this
> should best be done by the user agent, not by the introduction of
> extraneous characters to the content.

...and as I responded to Patrick, I don't disagree. However, in the here
and now...

> However, some text-only browsers
> don't provide this separation. My own *opinion* would be that the
> overhead
> of the brackets and corruption (perhaps this is not the best word) of
> the
> content provides a negative impact that outweighs the benefit the
> brackets
> might give to those using text-only browsers (a very small population
> - I
> would guess much smaller than the screen reader population).

While I hear what you are saying, how is this a "negative impact" on screen
reader users? From a conceptual perspective, it in fact *does* signal a
"chunk" different from the others. I've heard (as I sure you have) screen
readers float right past alt text (especially when the image is placed
in-line with, say, a paragraph) without the user realizing that what they
actually heard *was* alt text. Again, most regular screen reader users that
I have encountered tend to "tool up" their software to read faster than the
default install, as well as at a far less verbosity than the default.
However, I would be very interested to hear from actual users on this...

>
> While the inclusion of "photo", "painting", or "icon" might be
> appropriate
> in cases where it is important for the user to know the nature of the
> image (for instance, a photo as compared to a painting), I think most
> people would argue against identifying images as images within
> alternative
> text. And I can't think of any case where the word "image" would be
> appropriate within alternative text. The purpose of alternative text
> is to
> provide an alternative to the content and/or function of the image,
> not to
> identify the presence of an image.

Again, it's a question of balance. Does specifically identifying a page
element as an "image" impact negatively on screen reader users? While being
specific: [Painting - The Mona Lisa] is definitely more useful to all than a
generic: [Image - the painting of the Mona Lisa], in both instances I
suggest that we are providing more than just alt="The Mona Lisa". Is this
good or bad? I don't have a definitive answer; don't have any hard data to
base an answer on. But if we are to consider the impact on cognitive load
issues, I would suggest that providing this type of information *could* be
of benefit to that user group, and that providing the same information to
non-visual users is not that "harmful". Either way, I am glad we are having
this conversation, and again, screen reader users... Please give feedback.

>
> I do find it interesting that there is still such a divergent opinion
> of
> best practices for alternative text, the first principle of web
> accessibility. I'm close to completing an article that will present
> my own opinion of some best practices.

I agree, and perhaps one of the things that can emerge is some firmer
guidance on this topic from community consensus. Jump in everybody!

JF
---
John Foliot
Academic Technology Specialist - Online Accessibility
Stanford University
560 Escondido Mall
Meyer Library 181
Stanford, CA 94305-3093






From: zara
Date: Tue, Aug 22 2006 9:10AM
Subject: RE: FW: alt text size
← Previous message | Next message →

Many thanks to all who responded. Your comments as well as those received at French mailing list Accessiweb will be of great value to our work and will certainly help us document and analyse the various options brought up.

Thanks again,


Catherine




--
Catherine Roy, consultante

www.catherine-roy.net








From: John Foliot
Date: Tue, Aug 22 2006 12:00PM
Subject: RE: FW: alt text size
← Previous message | Next message →

zara wrote:
> Many thanks to all who responded. Your comments as well as those
> received at French mailing list Accessiweb will be of great value to
> our work and will certainly help us document and analyse the various
> options brought up.
>
> Thanks again,
>
>
> Catherine

Kat,

Hopefully you will share with us your synopsis when completed, as well as
your recommendation/standard.

JF
---
John Foliot
Academic Technology Specialist - Online Accessibility
Stanford University
560 Escondido Mall
Meyer Library 181
Stanford, CA 94305-3093






From: zara
Date: Wed, Aug 23 2006 10:10AM
Subject: RE: FW: alt text size
← Previous message | No next message

> Hopefully you will share with us your synopsis when completed, as well as
> your recommendation/standard.
>

This should not be a problem. As for the future standard itself, what I can say for the moment is that development on this comes notably from the adoption of article 26.5 in our province's disability legislation[1] in December 2004. This article stipulates that the government must adopt a policy ensuring that ministries and public organisations put in place reasonable accommodation measures permitting persons with disabilities to have access to documents, whatever their form, and services offered to the public. This policy must be in place by December 2006. This summer, our Ministry of Government Services and our province's equivalent of a Disability Issues Office started the process of developing a standard to support this obligation with regards to government information on the Web. So we are now working on the first proposal. This will then be analysed and discussed at the inter-ministerial level and after that, we will see how things go. It is planned that this process should be completed at the end of the present year.

Anyway, thanks for your interest.


Catherine
--

[1] <http://www2.publicationsduquebec.gouv.qc.ca/dynamicSearch/telecharge.php?type=2&;file=/E_20_1/E20_1.html> (in French)



--
Catherine Roy, consultante

www.catherine-roy.net