PDF tagging issues discovered by the PDF Association experts

[ronaldtse]

The PDF Association (who publishes the PDF spec) has reviewed Metanorma's use of PDF tagging and have the following comments, which we need to all resolve.

cc: @petervwyatt

We publish a "Tagged PDF Best Practice Guide: Syntax" (https://pdfa.org/resource/tagged-pdf-best-practice-guide-syntax/) which you should definitely be referring to.

• PDF tagging: poor tagging of ISO cover page metanorma-iso#999

• PDF tagging: ToC uses L instead of a TOCI metanorma-iso#1000

• PDF tagging: tables are structured wrongly, and include non-table content metanorma-iso#1001

• PDF tagging: Lists are getting tagged as paragraphs and not lists metanorma-iso#1002

• PDF tagging: Some bad semantics: footnote, P tags, etc metanorma-iso#1003

Also ran the generated PDFs through the Arlington PDF Model and a few other tools to analyse the PDF syntax:

• PDF issues: Metanorma uses PDF features from 1.4+ but it is tagged as %PDF-1.4 #202

• PDF issues: do not generate ProcSet array resources #203

• PDF issues: PDF Linearization data has bad errors #204

[petervwyatt]

@ronaldtse Are there any other "static" pages (like the ISO cover page) that could benefit from a professional review of their PDF logical structure and tags? It will be more efficient for someone to remediate a few pages all at once.

Maybe the 2nd page ISO disclaimer/copyright? Forward page? Back matter (final page)?
(NOTE: I haven't checked to see whether those pages have major issues, but a remediator will detect things I may miss)

[ronaldtse]

@petervwyatt I believe all our generated output would warrant a review. We could provide a sample ISO PDF file that incorporates most features and have that reviewed?

[petervwyatt]

That is also an option so long as the doc isn't too long - and especially if you have a fake/test ISO PDF that includes a fuller set of features that we may not be using in our own PDF ISO specs. Line numbering and "code" blocks come to mind, but there must be others... and table layout can be extremely variable.

[ronaldtse]

We typically use the ISO “rice” document to test the ISO layout, and we could definitely add line numbers and code blocks to it for testing purposes. The Rice document does include complex tables. Will check and revert.

[petervwyatt]

OK - please pass me an appropriate PDF when you're able to.

[petervwyatt]

I now have a volunteer remediation organization lined up and ready to make your test ISO publication PDF/UA-1 compliant.