A PSA on Creating Accessible PDFs

In a recent blog post, I discussed how CAPTCHA accessibility is the number-one accessibility issue faced by blind computer users. However, a close second to that is the accessibility (or lack thereof) of Portable Document Format (PDF) files -- but it's an challenge that can be solved, with the right tools and a little bit of time.

PDFs were created by Adobe Systems to be a cross-platform file format that all systems could render. A PDF file is meant to look as good on a computer screen as it would in print.

The trouble, from an accessibility point of view, was that PDF appeared as an image to text-to-speech programs which meant they couldn't parse any text. Optical character recognition (OCR), though not yet in the mainstream today, is only available as expensive commercial blind-specific products rather than as part of scanning software. If an image lacks colour contrast or is presented in an atypical font, the accuracy of OCR is further diminished.

So, for a long time, anything with a PDF extension was almost always unreadable to a majority of users requiring adaptive technologies. And considering the ubiquity of this format, it was frustrating that no alternative formats were available.

To its credit, Adobe is coming around on accessibility. Its Read Aloud feature can be handy, though it can't process images either. It has a mechanism for tagging PDF files. Read Aloud can also be configured to attempt to display untagged files in this format in a readable and logical order -- if it detects the use of screen reading or magnification software. As a result, many more PDF files are readable today.

But it seems that few know of, or are choosing to implement a simple process thank can make PDFs accessible from the beginning. As most PDF files are first created in a word-processing program and then later converted to the PDF format, creating an accessible PDF starts with first creating an accessible source document.

Creating an accessible document in a word processor is not a challenge in itself, but it does require using that program's mark-up features. This creates the file's tags – or the mark-up analogous to HTML tags -- that serve to inform an application in the manner and sequence that a file should be displayed.

For example, using Microsoft Word, lists, headings (which should follow a logical structure), and any other formatted text should be created using styles and appropriate colour contrast. You also want to ensure that all link and image text is descriptive. For images, this means placing descriptive text in the "title" and "description" fields in Word (as some screen readers will render one field and not the other when converting to other formats). And while table column headers should be specified, it's not always possible. For example, using Microsoft Word, row headers cannot be indicated in the same way.

The next trick to producing an accessible PDF, once all styles and alternative text have been applied, is to convert it properly. Though one would assume simply selecting to "print as a PDF" would see all tags and formatting preserved, it's not the case. Only by choosing to save the file as a PDF, or by using an add-in specific to Adobe, allows the conversion to work.

Though I don't intend this to be a specific endorsement of a software package, my experience is that Adobe Acrobat is one of the few software packages that can produce fully tagged PDF files.

The final step in creating an accessible PDF is to open the file in Adobe Acrobat professional and touch up any tagging errors the software may have made. This is a simple matter of using the menu system.

While not an exhaustive how-to guide, I want to stress how fairly straightforward creating accessible PDFs can be. As with your company website content, the accessibility -- or lack therof -- of other digital materials can reflect on your business as a whole.

After all, your message is important. It would be a shame to have it lost in digital translation.