PDF crosswalk and techniques

[rickj]

When discussing the EPUB and ONIX workflow, some publishers have asked about PDF. Specifically will the W3C be producing a crosswalk and a techniques document for those accessibility features supported by PDF, so that the appropriate statements in the principles document can be applied.

[DuffJohnson]

The PDF Association (pdfa.org) is currently working on several initiatives with direct connection to this question:

• The PDF/UA Processor LWG is studying the various accessibility API role mappings for HTML elements and WAI-ARIA (and DPub) attributes to map these features to their functional equivalents in PDF (tags, attributes, properties, etc.). The goal is to develop a new AAM for PDF under W3C.
• The PDF Accessibility LWG is developing formal techniques for accessible PDF. We expect the first set of these Techniques to be released in March / April of this year. An example of this work may be found here.

There are several other initiatives within the PDF Association on related subjects, these are summarized in this article. Other resources may be found on the PDF Association's landing page for accessibility.

Feel free to get in touch for more information.

[gregoriopellegrino]

Thanks @DuffJohnson. Are you also planning to embed accessibility metadata in PDFs?

[clapierre]

Yes thank you @DuffJohnson. You mention DPUB attributes, would this also include doc-header and doc-footer? Those are two ARIA role's introduced that needs implementations, and I would think this would be something PDF's would use more so than EPUBs.

[DuffJohnson]

@gregoriopellegrino WTPDF and PDF/UA both include "conformance claim" metadata in the file as XMP; that is, formal conformance with the spec requires inclusion of the claim.

WTPDF uses the PDF Declarations model developed by the PDF Association, while PDF/UA's model is the same as is used for all ISO-standardized subsets of PDF technology.

The list of industry-recognized Declarations includes 3 defined for "accessibility validation" in reference to WTPDF and PDF/UA. On request the PDF Association would - I'm sure - be happy to add validation declarations for WCAG as well.

Each such Declaration includes recommended fields for "claimBy, claimDate" and optional fields for "claimCredentials" and "claimReport".

[DuffJohnson]

@clapierre PDF 2.0, WTPDF and PDF/UA don't prescribe (or restrict) any specific roles - they merely open the doors for people to use what makes sense to them.

[clapierre]

@DuffJohnson So I assume that claimBy, claimCredentials, and claimReport map to EPUB accessibility metadata of

property="a11y:certifiedBy" property="a11y:certifierCredential" property="a11y:certifierReport"

as defined in EPUB Accessibility 1.1: Conformance and Discoverability Requirements for EPUB publications

[DuffJohnson]

We've not so-stated, but that is an entirely reasonable assumption, yes.

[avneeshsingh]

Thanks @DuffJohnson for making us aware of accessibility metadata initiatives in PDF. While developing embedded metadata for EPUB 3, we realized that a lot of metadata properties and values that we need are very well defined in schema.org accessibility metadata. Therefore, we decided to incorporate the metadata from schema.org to the extent possible, and invent the new metadata properties in EPUB 3 only if they were not available in schema.org.

If the same approach can be followed in PDF, then it will greatly reduce your work, and will make it much more easy to develop the cross walk that @rickj mentioned.

[DuffJohnson]

Thanks for making this point. I've added this question to our agenda, and will update this issue with the resolution.

[rickj]

@DuffJohnson just wondering if there has been any update from discussions you were a part of