Should alt text be used to paint a thousand words?

We’ve all been told that alt text is an essential part of web accessibility, but how much detail do we actually need to include and who should do it?

There’s been some discussion over on the DC-Accessibility JISCMail Discussion List (February 2007, “Not Accessible or Adaptable”) about a lot of issues, including whether alt text should always be added to an image. One contributor to the discussion gave a link to a slideshow of dance photographs, where:

“the author refused to label the images with text… his argument being that the photographers images capture and demonstrate an emotional experience, and that whilst text can perform the same expression, he’s not the person to annotate them.”

The photographs in question are various stills from dance rehearsals and performances.  There is no accompanying text of any kind, but most people would probably recognise that the people in the photos were involved in some sort of dance medium from the clothes being worn, the environment, and from the positions of the bodies.  However, unless one knows the language of dance or the context in which the dance is being performed, the photos may have no further meaning – and could therefore be inaccessible to some people.

This actually brings up several issues:

1. How can one describe an image that expresses emotion or abstract concepts?

2.  If such concepts can be described, who should be responsible (and have the capability) for doing so?

3.  Where does alt text fit into all this?

1. Describing Emotion and Abstract Concepts

So is it possible to extract emotional and abstract meanings and describe them for people who do not have a concept or understanding of such areas?  The Dayton Art Institute Access Art Website has attempted to do so.  For each artwork on the Access Art website, there is an image, a section on the artwork in context, comments by the Art Director (including an audio commentary) and a description of the artwork. Each section is no more than a couple of paragraphs. For example, the description of Frishmuth’s “Joy of the Waters” has attempted to put across abstract concepts such as the mood of the statue:

“The girl’s springing, energetic step, joyful expression, and animated hair create an exuberant mood and suggest that she may be a water sprite.” (Marianne Richter, Dayton Art Institute)

This helps make the artwork become more accessible for visually impaired people and for people who do not know the language of art. 

2. Responsibility for Describing Images

The people best qualified to describe a visual resource are probably the people who have decided it should be included in the first place.  For example, someone with archaeological experience is probably best placed to describe an image of a stone tool, whilst a geography tutor may be the most suitable person to describe a meteorological image from a satellite put onto the university’s VLE (Virtual Learning Environment).

The descriptions used will also differ depending on the image’s intended audience.  A museum generally has a wide public audience with many different levels of understanding and access requirements, whilst a Geography department may only have a small number of students at a fairly high level of understanding. 

So, unless the photographer in the quote above, is also versed in the language of dance, he is unlikely to be able to describe the dance photos he has taken.  Even if he were, he would also need to be aware of the level at which they needed to be pitched in terms of language, description, and audience. 

3. Use of alt Text

So where does alt text fit into all this?  The W3C (World Wide Web Consortium) WCAG (Web Content Accessibility Guidelines) recommends providing:

“…a text equivalent for every non-text element (e.g., via “alt”, “longdesc”,  or in element content)… For complex content (e.g., a chart) where the “alt” text does not provide a complete text equivalent, provide an additional description using, for example, “longdesc” with IMG or FRAME, a link inside an OBJECT element, or a description link.”

Therefore, alt text should be used for every image (even empty alt text should be used for spacers and decorative images), but it should only provide a brief text description of the image – the Guidelines on ALT Texts in IMG Elements recommends no more than 50 characters.  Longer descriptions of an image, such as those describing complex images, emotions, or abstract concepts, should not be included as alt text, but should either be attached as a separate link (perhaps using the longdesc or d-link elements) or added next to the image. 

Alt text can also be different for different audiences and purposes (see WebAIM’s Communicating the Purpose of the Graphic) and does not necessarily need to be completed by experts.  However, although the photos of the dancers should have had alt text, they may well have needed someone with a knowledge of dance to add it.  Basic alt text, such as “photo of dance students” could have been added by anyone but would there be any benefit to seeing roughly the same alt text added to over 80 images?  A choreographer would be capable of adding more informative alt text, such as stating the dance step or intention, e.g. “photo of a dancer in fifth position”, particularly where the intended audience was other dancers or dance students. 

Alt text is a requirement under the WCAG guidelines, but it shouldn’t be used to describe an image in a thousand words – these have to be written elsewhere.

9 thoughts on “Should alt text be used to paint a thousand words?

  1. excellent piece Sharon which captures much of the essence of the whole accessibility debate, not just descriptions of images on Websites. what is it for, who is it for and who should do it…
    to address the questions you raised:
    1. How can one describe an image that expresses emotion or abstract concepts?
    Answer: in a learning context the description is there to describe the purpose of the image rather than necessarily its exact features. Descriptions like the images themselves are contextual. is understanding the emotion essential to completing the objective? only if it is do you need to describe it.
    2.  If such concepts can be described, who should be responsible (and have the capability) for doing so?
    Answer: following from #1 above. it is essential that the person selecting the image is the one who describes it. It is only they who can explain why it was chosen and who can describe it in such a way as to not present a barrier to achieving the objective. It is no good having a description of an image which says ‘happy girl’ or ‘girl in a green dress’ if the objective is to assess the mood of the girl. the author has the responsibility and by the nature of having selected the image must have the ability to describe its purpose.

    3.  Where does alt text fit into all this?
    Answer: ‘alt’ is simply a technical tool, one tag in one language, it isn’t the door to paradise any more than a single textual description is the key to that door. alt can be used to describe the purpose of an image, or the image itself but it isn’t the only way of doing either, nor would we want it to be. alt has many disadvantages as well as much in its favour. As long ago as 1999 we took the decision to remove alt tags and place the descriptions in the body of text in courses both because they were useful to all students and also because it made writing them much easier for expert authors who struggled to get to grips with the ‘extra’ alt tags.

  2. Wonderful piece Sharon.
    Adrian, I agree with your comments about purpose.

    I think there are still some very interesting questions around the provision of alternatives (not just alternatives for visual) when there is emotional or artistic purpose. My guess if you asked an artist to define what is art is that they would say its completely dependant on context. Art that makes sense in one context (say 19th century France) would not have the same effect in a different one (say modern Britain). That notion does not stop at cultures but extends to individuals too. Perception and understanding are not static and passive they are active. We participate in them and we are all very different. Also a painting or a poem or a piece of music is much more than its content and maybe much more than its purpose. A description of an emotion is not the same as the emotion. This is what makes art, poetry, education (and in fact even life) such fun – its guesswork. We simply don’t know how something will be perceived by anyone at all, so we guess, and in that approximate guesswork is the possibilty that we are completely wrong or very innacurate. But we try and that’s the point.

    We must do our best to provide materials in ways that best communicate to whoever the audience may be. A different experience is still worthwhile.

    I am minded of the work my cycling group is doing to support cycling for blind people by providing tandem riding. Of course this is a totally different experience to solo cycling and even to sighted tandem riding but its still a wonderful experience.

    Art is I think a communication. Each of us will have a different emotional response. That doesn’t matter and doesn’t invalidate the experience. The communication is between the author (who had the purpose) and the recipient (who has the perception mechanisms whatever they are). It *isn’t* about dead material.

    What I’m saying is that purpose is all, because its an indispensable part of communication and since its communication we are trying to do we should try to do that and that requires providing our best efforts to meet perception requirements that we may not be used to providing because they are different to what we are used to.

    I don’t believe in art for the elite, which is what it would be if we didn’t try to communicate with everyone that was in the audience.


  3. Pingback: Styling alt tags at Blog

  4. I think that arguments such as “The photo contains emotional content that I’m not qualified to relate in words” are complete cop outs.

    These authors are forgetting the purpose of the alt attribute: to convey some information about the image (other than “PIC01837463.JPG”) to users who can not perceive the visusal content of the image, such as the googlebot, and users with vision disabilities.

    In the case of these dance images, the googlebot just needs some content it can show in search results “photo of a dance performance”. This same alt text works just as well for a person who is blind.

    The blind user is a mature, thinking human being, who probablly is very well aware that he or she will never get the same content from an image that a sighted user would. However, the alt text is there so this user can put the images into the context of the rest of the site: She now knows there are 14 images of dancing, rather than a whole bunch of JAWS-babble “img: PIC0034738.JPG”.

    Even short alt text provides context, and improves the sound of the site to users of screen readers. It’s completely unnecessary to try to convey the exact same content. It is necessary to make a bit of effort to convey that there are in fact images (don’t use alt=”” where inapropriate), and what their subject matter is.

  5. Pingback: Avoiding the Gray Areas - Curb Cut

  6. Sharon,
    Interesting, however my blog is approaching a similar issue from a digital imagery in general and fine art in particular argument. The main consensus of my research to-date seems to agree, that an image requires a caption / title, as an aid to understanding that image and that is the point, the artist does not need to explain the meaning of the image. It could be argued that the onus is with the artist to give guidance to possible meanings which is for the viewer to explore.

  7. Pingback: Bilder, User, Social News und Social Communities | Online Marketing Blog

Leave a Reply to Colin Lieberman Cancel reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>