Home

Integrating accessibility into PDF documents: Don't panic!

Publication Date: 
June 18, 2008
Expiration Date: 
June 18, 2011
Erika Donald, OCIO—Security, Privacy, and Policy
Weight: 
0
Body Text: 

Sean Keegan, web accessibility instructor at the High Tech Center Training Unit of the California Community Colleges, opened his "Integrating Accessibility into PDF Documents" presentation at UC Berkeley's joint Webnet and WebAccess meeting on May 6, by admitting he had an alternate working title for the presentation: Don't Panic.

"Really though, it's not a hard process," he insisted. By the end of his 90-minute, highly informative presentation, the process of improving document accessibility in PDFs felt like no sweat at all.

Three PDF flavors: Image only, searchable, or tagged

Sean began by defining the "three flavors of a PDF" with respect to accessibility:

  • the image-only PDF which is only a picture of the content — i.e., any text in it is just an image of the text that is not searchable
  • the searchable PDF which has actual text content yet is lacking an embedded document structure to guide accessibility reader programs
  • the tagged PDF which has an embedded structure that allows accessibility reader technology to identify the logical order

He delivered lessons for the "easy method" for creating PDF documents (requiring MS Office 2000 or later and Adobe Acrobat 8), and the "less easy method" when you are starting with an original document that is not navigable or is in an earlier PDF format without tags.

"What you end up doing when building a foundation for accessibility," he said, emphasizing a crucial point of the presentation, "is making the PDF documents more usable for everyone, whether they use assistive technologies or not."

Mac users: A little bit of extra work

"Anyone here use a Mac?", Sean asked in jest as many in the audience raised their hands. Even though built-in tagged PDF is not available for the Mac yet, "all is not lost," Sean joked. He then gave a detailed explanation of the "little bit of extra work" required for Macintosh versions of Adobe Acrobat, which involves running the "Add Tag to Document" feature to generate a tagged PDF document. But still, the news was positive: "Most searchable-level PDFs coming out of the Mac platform work really well. But he cautioned Mac users, "remember to save and save often. There's no 'undo' command in the tag-adding software."

Reading order is critical: Does it make sense?

"Reading order is critical" he stressed. "Too often people throw assistive technologies at a PDF and say 'just read the content'." The content may all be there, but does it read in a way that makes sense? For example, text boxes, columns, and other structures may cause a text reading program to pull in sections of a document out of sequence.

To avoid the whole "tedious" and often ineffectual process of adding tags, Sean's recommendation: Start with your document editor, "do a quick and dirty assessment" of the document to make sure things are in the right place, then export it to an acessible text document or put it in plain text to visually check the reading order.

In answer to questions from the audience about generating PDF documents from PowerPoint and Excel, he was equally helpful. PowerPoint has an option to generate tagged PDF documents that are correctly tagged to retain the reading order, and Excel files may require only a "slight touch-up" after conversion to PDF. But, he said, "assistive technologies interact directly with Excel just fine" obviating the need to generate a PDF version of the file for your website at all.

As he wrapped up his presentation, Sean gave one final overriding recommendation to the audience. "Make your web pages primarily in web format and then, only as a second-choice option, post PDF versions of documents — and try to make them as accessible as possible." In other words, build a web-based foundation for access from the ground up.

Presentation resources

Webcast and podcast. A webcast and podcast of Sean Keegan's presentation, produced by UC Berkeley's Educational Technology Services (ETS), are available on:

The webcast is also available on:

Transcript. A transcript of the presentation, produced by the Realtime Captioning unit of UC Berkeley's Disabled Students' Program is available at: http://webnet.berkeley.edu/docs/PDF_Accessibility_transcript_05062008.doc.

Slide presentation. The PowerPoint slides from Sean's presentation is available at: http://webnet.berkeley.edu/docs/Accessible_PDF_documents_05062008.ppt.