WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: Question about image in the alt attribute

for

From: Jonathan Metz
Date: Aug 1, 2014 4:43PM

On 8/1/14, 4:33 PM, "Jukka K. Korpela" < <EMAIL REMOVED> > wrote:

>Attributes are not tags.


I'm finding it irritating that you continue to point out an error I made
by mistake — after recognizing you understood what I meant — despite not
adding value to the conversation.


>All HTML specifications, and also HTML5 drafts, define the alt attribute
>as specifying a textual replacement for the image, for use when the
>image is not displayed. They are partly rather sloppy in their wordings,
>also referring to "descriptions" rather than replacements, but the
>definitions proper (as opposite to annotations) are rather clear.


My main argument has been to focus on the way WCAG defines it because HTML
specifications are intended for more than just one purpose. They have a
wide range of applications, one of which is accessibility and we should
focus on instructions that deal with that instead.

It makes sense for HTML specifications to define the alt attribute this
way, because HTML standards discuss the relationship between a language
and technology.

On the other hand, the Web Content Accessibility Guidelines discuss a
relationship between technology and users when implementing said standard.
In this regard, WCAG defines a fundamental means of describing something
differently in order to be understood by a variety of different methods.


>Most importantly, we need to decide whether an alt attribute is written
>to act as a replacement (for use when the image is not seen at all) or
>as an annotation/commentary/description. You can't have it both ways, in
>general.


I *think* I agree with you here. I do not believe that text alternatives
should refer to a relationship between the content and the non-text
element, but I don't think I've said that during this conversation.


>In the case of <img>, the HTML language provides exactly two alternative
>ways: an image and a piece of text. How the text (when it is rendered)
>is presented to the user will depend on the browser.

It sounds like you are trying to explain the relationship between
technology and code. I am more interested in determining the purpose of
alternate text being for more than one disability, which I surmised based
on your original response to Bevi about how her understanding of the alt
tag is not the intended purpose:

Bevi:
>Please keep in mind that WCAG's standards/guidelines for Alt-text aren't
>only for those who are fully blind. Partially-sighted and low-vision users
>rely on Alt-text, also. People with various dyslexia as well as
>autism/Asperger's also benefit from Alt-text on graphics because it can
>help
>reinforce their perception of the graphic.

Yucca (I'm sorry, is it Yucca or Jukka?):
>That's not what the alt attribute is for. It is to be presented when the
>image is not displayed. It is ALTernative.

People in this field have a tendency to be disability- or
technology-specific at the detriment of furthering a goal of inclusion.

>>4.7.1.1.1 Examples of scenarios where users benefit from text
>>alternatives
>>for images
>>
>>* They have a very slow connection and are browsing with images disabled.
>>* They have a vision impairment and use text to speech software.
>>* They have a cognitive impairment and use text to speech software.
>>* They are using a text-only browser.
>>* They are listening to the page being read out by a voice Web browser.
>>* They have images disabled to save on download costs.
>>* They have problems loading images or the source of an image is wrong.
>
>These all mean that the image is not seen.


No, that is not accurate.

4.7.1.1, Requirements for providing text to act as an alternative for
images, doesn't state that alt text is provided when the image is not seen
— only that there is an *alternative*. You are personally applying a
reason of using an alternative, but HTML5 does not state that images are
in fact unseen.

If this is was the case, why would there ever be alternate text for videos
(captioning)?

>>I don't think I understand you here. Are you saying that if a photo is
>>included, there is no way to describe what is going on in the photo?
>
>Yes. You might be able to describe part of its content somehow, but it
>is an illusion to think that one can write adequate textual replacements
>for photos, except in rather trivial cases. Pick up any photo from your
>personal photo gallery and think how you would describe its content over
>the phone. Of course you can tell that it is a photo of uncle Joe, but
>that would apply to any photo of uncle Joe, so it can hardly contain the
>information content of a particular photo.

That's not the purpose of alternate text though. There needs to be enough
information for someone to grasp an understanding of what is going on. If
I were to tell you that I have a photo of President Obama, to compensate
for a lack of having never seen this man, you would be able to
conceptualize an understanding of the image based on your own senses. For
example, one might hear him speaking when thinking about him.

Keep in mind that while there are many of us out there who think in terms
of images, and that is actually how we communicate with the outside world,
we may still rely on authors to provide us with more information for their
choice of image.

For example, WebAIM has an excellent example of this in practice
<http://webaim.org/articles/cognitive/#reading>;. When I came across the
text:

"Tob eornot obe"

I stared at it for a really long time, but without the image I was having
trouble figuring it out.

If I was still having problems figuring it out though, I could Inspect the
image and sure enough there is alt="portrait of William Shakespeare"
included. I might not have ever seen a picture of Shakespeare, but I could
acknowledge a relationship between the image and "To be or Not to Be"
because it provided related context.


>When an image is used as a list bullet, I would use alt="item" or
>alt="bullet" or something like that. Wouldn't you? Or would you explain
>all the details of the bullet image?

I wouldn't. HTML5 (thankfully) returned type (among others). With HTML 4,
we relied on CSS instead.

>>Still, if you came
>>across a subtle background of a seashell pattern on a web site about
>>coral
>>plant life, would you really want to be notified that the designer used a
>>repeating clip art graphic?
>
>No, not any more than I want the random words read aloud. As I wrote,
>alt="" is adequate, but definitely not because the image is purely
>decorative
>
>>>But it's an image for which alt="" is adequate, on the simple grounds
>>>that this is the only no-nonsense approach. There is no way (or at least
>>>I cannot see a way) to present the same information in a non-visual way
>>>in a similarly non-disruptive manner.
>>
>>So… You're agreeing that this was a good example then?
>
>No, it's a poor example, because it presents the case as an example of
>purely decorative image, when it is no way near purely decorative. A
>nice image of a beautiful butterfly is purely decorative on a page that
>has absolutely nothing to do with butterflies. On a page about insects,
>it is a topical image. On a page about butterflies, even more so.

If there is an image of a butterfly on a page about insects, I would argue
it's filler if the butterfly is never mentioned on the page.

If the content was about chaos theory, then I certainly wouldn't provide
alt text to a pretty picture of a butterfly or a wheelchair simply because
it had something to do with Ashton Kutcher.

Users should be mindful about the use of alternate text, and using it
because it *sort of* makes sense in theory isn't a good reason to add it.
If the conversation is about Senior Health, and you provide alt text
referring to a portrait of a septuagenarian, that isn't helpful. However,
if you are including alternate descriptions of a 70 year old performing
yoga, there is merit due to the relationship between Seniors and Health.


>You draw such conclusions at your own risk.


Therein lies the rub though. It isn't a risk that I am making because of a
bias, it's established based on how the author decided the non-text
element was not worthy of the content that it is sitting next to.


>The alt text isn't even
>meant to be seen when the image is displayed.


Again, this is not accurate. "Programmatically associated text is text
whose location can be programmatically determined from the non-text
content"
(<http://www.w3.org/TR/UNDERSTANDING-WCAG20/text-equiv-all.html#non-text-co
ntentdef>) This implies a means of recognizing text.


>There are myriads of
>reasons, some of them rather good, why an image might have alt="" even
>though seeing it helps in understanding the page content. If there is no
>way to help a person who does not see the image the same way as the
>image may help a person who sees it, it is much better to use alt=""
>than to write something that does not help.

In the past, I've heard similar statements made from programmers who argue
that the camera shouldn't be exposed to accessibility features in mobile
apps since it the camera is meant for capturing visual content. Or perhaps
how video games are rather visual, there could be no reason why a blind
person would ever play.

Proper and thoughtful addition of alternate text for relevant content can
be helpful for people in ways you might not have thought of.


>In order to explain how a photo relates to the rest of the content, or
>vice versa, normal text should be used, in running text or as an image
>caption. After all, that text would be meant for people who see the
>image, not to those that don't.


In my opinion, this is a fundamentally limited and potentially damaging
opinion regarding the purpose of accessibility.

Jon