Alt? Longdesc? Title? What goes where, especially for image-heavy sites?

Question

So I've been trying to make my site as accessible as possible (for non-JavaScript users, web crawlers, screen readers, etc), and I hit a large snag.

The site I'm developing is very image heavy. (I draw stupid stuff in my spare time and the site is a sort of a "showcase".) Each page in the site has a single image (or multiple, if context is needed or the image is several panels) that is usually accompanied by a single caption underneath. So when I got to the point where I had to add accessibility options, I wasn't sure what to do. For example, let's say I had an image of a man eating an apple and the joke was there's a worm in it. So the first thing I did was add a ludicrously descriptive alt text since it was the only way to provide an "equivalent" for screen readers.

<img src="appleeat.png" alt="A man bites into an apple. As it turns out, there's a worm in it!">

This seemed okay at first, but then things went awry really quickly when I suddenly found myself needing to add alt text up to 300 characters (!!!) for the more elaborate jokes and images. Not only that, but apparently alt text isn't for descriptions anyway.

So I'm having these problems:

1. Alt text is for equivalents, not descriptions. However, the only way to provide an equivalent in this case is to provide a description.
2. longdesc seems to be more suited for this feat, but longdesc is not supported by any browser (at least, according to W3Schools) even though it does seem to be in the HTML5 specification.
3. <figcaption> seems to be the way to go, but it ends up displaying text underneath the image in question, which is definitely not very appealing for my site, especially if I want to add additional captions and context in <p> tags or something.

So what do I do? What would I put, and where? I'm totally stumped on this, and frankly I'm not certain that making a site based entirely on images accessible to people who can't see very well is a good idea.

Answer 1

Most people now use a figcaption as a caption under an image in an article, or for a short image description, like so...

<figure aria-labelledby="myimage1" >
    <img id="image1" src="myimage.jpg" width="100" height="100" alt="image:My Image" title="Image of Something Cool" />
    <figcaption id="myimage1">This is My Image Caption!</cite></figcaption>
</figure>

However....since you are associating a larger description with your image, and it is not needing a caption, try this...

<div role="img" aria-labelledby="myimage2">
  <img id="image2" src="myimage.jpg" width="100" height="100" alt="image:My Image" title="Image of Something Cool" />
  <p id="myimage2">Larger description of my image...</p>
</div>

This is also compatible with WAI-ARIA accessibility standards, which will read and associate your container "div" and your "p" description with the image. Notice the new attribute, role=img. Since the div, unlike figure, is not normally associated with an image, this helps the screen reader identify the div container as part of the image.

ADDITIONAL

Note that the alt attribute should be a text replacement for an image if the image never appears, a user agent cannot read images, or the image is slow to download. I like to show this info in a grey background image box in case a image breaks and a user needs to know what is missing. Its designed to tell a user what the image represents in a simple, short text string.

The title attribute on images is an important rollover box that gives additional image information, image map feedback, or where it would link to when the image is a clickable hyperlink. Some designers frown on the title attribute but I use it all the time in a lot of my HTML to add rich info to all my interactive and visual elements.

Answer 2

The alt attribute is for alternative text, i.e. textual replacement for an image, so in the example, it is adequate if it reasonably tells the same story as the image. In reality, most images cannot have texts that are full “alternatives” or “replacements”; it’s usually a matter of capturing some of the most essential message, if possible.

An alt text can be of any length. The statement that alt texts should not be descriptions does not mean that it cannot be detailed if needed. The point is that there are too many descriptions that say something about an image without conveying its message (like “big red bullet” or “A man in a canoe”).

The longdesc attribute is supported by some software but highly debated and not part of W3C HTML5 CR but being developed as an independent “extension”.

The figcaption element is for captions presented along with an image. It does not address the issue of alternative text at all. It is meant to be presented to the user, whether he sees the image or not.

Answer 3

If you want to provide complex alternative text (where "complex" could mean: containing lists, tables, audio/video etc.), you could either use an img element with a longdesc attribute, or an object element instead of img.

longdesc

In HTML 4.01, longdesc is part of the specification. In HTML5, it was removed, and is now developed as an HTML5 extension (but it’s currently only a Working Draft from 2013 Update: it became a W3C Recommendation).

Problem: You would either have to include the content on the same page, but then you shouldn’t visually hide it (as not only screen reader users may want to access this content). Or you would have to add separate pages for the content (but then search engines won’t relate this content to the image, i.e., you miss on ranking potential).

object

The object element can be used for any kind of media. Its content is the fallback content, which allows you to use markup for the alternative text.

<object data="appleeat.png" type="image/png">
  <!-- the alternative content goes here -->
  <p>A man bites into an apple. As it turns out, there's a worm in it!</p>
</object>

Answer 4

Here’s some discussion of alt text that might apply to your situation:

• http://webaim.org/techniques/alttext/#context.

Although I agree with @Jukka that long alt text isn’t a problem in itself. It’s true that “alt text isn't for descriptions”, but only because alt text is for equivalent content. If the equivalent content to a particular image is a description of it, then that’s fine.

It’s just that for lots of images used on the web, that’s not true. For example, the best alt text for the Stack Overflow logo at the top of this page wouldn’t be

The Stack Overflow logo: an abstract in-tray overflowing with paper, followed by the word “stack”, and in bold, the word “overflow”

It would just be “Stack Overflow”, because that provides an equivalent experience for a partially-sighted person (i.e. it tells them which website they’re on).

But if your images are cartoons, I guess writing alt text for them is a bit like doing audio descriptions for movies. You need to be more descriptive if you want to provide an equivalent experience. It’s certainly a challenge, especially if you don’t have experience, or partially-sighted users to discuss it with.

Joe Clark, in his book “Building Accessible Websites”, discussed describing images:

• http://joeclark.org/book/sashay/serialization/Chapter06.html#h2-2290

And even providing alt text for comic strips, although his example is for dialogue-heavy strips, and I’m not sure what sort of experience it provides in today’s screen readers:

• http://joeclark.org/book/sashay/serialization/Chapter06.html#h2-4240