Accessibility tagging is moving to a new level as innovation is added into the ability to auto tag documents for PDF or HTML5. Algorithms and future artificial intelligence will bring document tagging and document structure interrogation to a whole new level. The days required for manual tagging will be drastically reduced and structured publications will be tagged instantly with 100% accuracy. Of course, we also should have flying cars and robots that cook our meals!
Auto tagging for accessibility will be a reality when authors and organizations can rely on structured formatting i.e. a logical hierarchy within a given document. This includes the proper use of fonts, document structure and content. Document design must have consistency so that algorithms can expose content and look holistically at a document in order for it to be properly tagged.
Millions of documents are created daily. While the file may convey the proper information, the reality is that not everyone has the necessary training to publish documents, accessible or otherwise. As the demand for document accessibility increases, some organizations are designing documents with accessibility in mind (as opposed to some of the more recent white space and photo art designs). Good document design and consistent use of font types for headers, titles, figures, and other formats will easily allow for auto tagging.
We know that there are many documents stored as PDFs in archives. Some of these are external facing documents while others are for internal organizational consumption. Since documents can be created and published by anyone within an organization, corporate templates and single publishing workflows are not always standardized or used consistently. As a result, documents cannot always achieve 100% accuracy in auto tagging.
How Can Auto Tagging Become a Panacea?
An easy answer is to standardize compatible templates to accessibility. But we know this is not often done consistently. As such, we need to focus on quality control and quality assurance. What can’t be tagged automatically to a 100% pass falls under quality assurance testing and post tagging remediation.
As with manual remediation, a significant portion of the time required to tag a document can be spent doing quality control. Auto tagging greatly reduces the amount of time needed for tagging, thereby reducing the overall cost. The process is much more efficient, but the importance of accuracy and legislative compliance remains.
During the quality control phase we use both machine and human verification. PAC 2.0 and Adobe’s Accessibility Checker are two very good options that will catch many of the issues affecting accessibility. As good as automated tools can be, it’s important to have someone manually check the tags and read order of the document. The other element of human verification is to test the files using the assistive technology that clients actually use. Screen readers such as JAWS and NVDA, as well as an iPhone running VoiceOver are examples of the technology that can be used to confirm that files have been made accessible. Anyone involved in this kind of testing needs to have a thorough knowledge of the keyboard commands and gestures that a user would employ, or else the results from the quality control process will not be reliable. Note that Adobe Read Aloud was never intended to function as a screen reader and should not be used to determine compliance with accessibility standards. Although both machine and human verification are critical for quality control, auto tagging will dramatically decrease the time required for both of these processes.
It’s not too late to attend next week’s webinar on Testing Accessible Documents, being held at 1:00 ET on Tuesday, July 26. Sign up here, or if you miss it, check our website to see what other webinars we have planned and to access previous webinar recordings.