I am talking about a site in which certain images contain mission-critical information. I was wondering: would a solution in which a PDF file is replacing an image description (and ALT) ever be considered accessible according to the WCAG 2.1 standard?

Would the acceptability depend on the accessibility of the PDF itself? If the PDF contains a screen-readable version, it will functionally be possible for the user to get the information, if not convenient. However, if the PDF fails the accessibility check, then it would naturally disqualify.

What do you think?


Please note that success criterion 1.1.1 uses the following definition of text alternative (emphasis added):

Text that is programmatically associated with non-text content or referred to from text that is programmatically associated with non-text content. Programmatically associated text is text whose location can be programmatically determined from the non-text content.

Based on the question whether a PDF file is acceptable, I assume that the image requires a long description. However, even in that case, the alt attribute cannot be omitted; it should briefly identify the image and possibly mention where a longer description is available.

The longdesc attribute is one way of linking to a longer description (in addition to, not instead of, the alt attribute). Note that the longdesc attribute may contain a link to a different page or to a different location on the same page (using a fragment identifier).

Due to poor support for longdesc, there are a number of alternatives, such as using aria-describedby or linking to a long description from a figcaption element (which would be programmatically associated with the figure element enclosing the img element).

One of the best solutions I have seen so far is the following:

  • enclose the img element in a figure element,
  • add a figcaption element,
  • inside the figcaption element, use details element for the actual figure caption text and a summary element for the long description.

This allows you to have everything on the same page without cluttering the page for those who don't need to see the long description.

I would advise against making the long description available only in a PDF file. You might be able to argue that it meets the letter of WCAG 2.1 SC 1.1.1 if you link to it from a figcaption element, but as a certified auditor, I would never let this pass without a strong recommendation to use an HTML-based technique. Forcing screen reader users to open or download a PDF file for a long description does not constitute equal access, even if the PDF is accessible.

Linking to a PDF file instead of using an alt attribute would constitute an unequivocal violation of SC 1.1.1.

  • This seems very clear. In this case it should be noted that the image is more of a placeholder for PDF, a front page rather than the full document which could be multiple pages.
    – AlFin
    Oct 4 '20 at 6:23
  • @AlFin You said that "certain images contain mission-critical information", not that they are "placeholders for PDF". Please don't move the goalposts.
    – Tsundoku
    Oct 4 '20 at 12:11
  • Sorry for the imprecise initial phrasing. The relationship of the image and the PDF is complicated, as the image preview is usually enough for a (sighted) user to achieve their goal. So it is where their mission end even if they might seldom as get the same information as PDF maybe to archive it but generally the image suffices.
    – AlFin
    Oct 6 '20 at 4:37

You're probably covered at a very base level about the idea of providing alternate locations for very long image descriptions as covered in "G73: Providing a long description in another location with a link to it that is immediately adjacent to the non-text content", though this only talks about alternate HTML pages, if you're providing an accessible document I'm sure that's moderately sufficient.

In the spirit of great accessibility, though, I always ask myself "What would be an even better experience? How could we engage more of an audience?". To that question I recommend trying to provide that information directly on the HTML page or at the very least in another URL linked below the image. Think about a user using a screen reader: while opening a PDF and reading it seems easy to us, jumping between applications or document types is probably somewhat of a chore than having one format.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.