This one’s going to be controversial. I’m doing a lot of coding at the minute, sticking as best I can (naturally) to using DIVs, CSS and all that. It’s quite refreshing to realise that you can almost entirely eliminate imagery from the page code itself. But, of course, not always. And if you’re using inline images, you have to use ALT text, for accessibility and all that. Right? Well, maybe. I’m no longer convinced that it’s an absolute.
The W3C guidelines are pretty clear. Indeed, it’s actually checkpoint 1.1. ‘Provide a text equivalent for every non-text element (e.g., via “alt”…’ But let’s bear in mind these guidelines were published eight and half years ago: the days of banner imagery, text in GIF files, and all that. Most of which, thankfully, is behind us.
Here’s the quandry. Let’s say you have an article about Joe Soap. You’ve got Joe Soap’s name in the headline, and Joe Soap’s name throughout the article. You decide to add a picture of Joe Soap to make your page look a bit prettier. It’s just any old picture of Joe Soap; it’s not directly relevant to the story. Do you:
- interrupt the screen reader’s passage through the text of the story, to say ‘it’s an image of Joe Soap’? (Like it’s going to be an image of something else?) or
- leave the ALT text blank?
My inclination is the latter. It’s surely better, in terms of usability generally, to make a judgement call, rather than relying on the absolute. Does the reader need to know what the image is? Is it essential / important / helpful to the reader’s experience? It may not be.
The Wikipedia page on the subject says: ‘While the use of meaningful alt text is necessary to comply with accessibility standards, and is good practice, sometimes an image is used for purely decorative purposes. In this case, one should use an empty alt attribute (alt=””).’ That sounds about right to me. But strictly speaking, that’s not what the W3C guidelines say… and that’s what we’re supposedly held to.
I’m of the view, no matter how obvious, that signposting stuff is worthwhile. So if it is an image of Joe Soap, call it out as such. If that means someone catching on screen reader gets information like “It’s an image of Joe Soap”, its up to them to process that. You can’t assume the criteria to which they’re evaluating the page. If it is there, tell them it is there. They may even be wondering about the page being text heavy.
Finally, like link text has improved to have some personality, I wouldn’t be surprised if ALT text followed suit. So expect to see “A particularly heroic portrait of Joe Soap” coming to a screen reader near you.
I sympathise with your thinking.
Last night at a bbc.co.uk consultation – it’s not too late to lodge your comments – I witnessed a teenager with learning disabilities and a visual impairment expertly navigating around bbc.co.uk.
It was amazing to see the Betsie/text-only option in full swing.
Where pictures had been inserted to illustrate the article, the wording of the ALT tab was often so good that it just became another explanatory sentence in the narrative – part of the sea of extra large yellow rows on a black background. And the beauty was that it was seamless. As if someone had actually checked through the pages in text-only mode and edited the captions to make sure it was as good as possible.
My conclusion would be that it’s important to make ALT text fit in with the main body of text around it. And it’s important not to add distraction by needless textual adornment of screen furniture UNLESS the user needs to know it’s there, or a search engine needs to be able to identify and index your image.