Autotag puts all links inside reference tags

[Oberlin Webteam]

Recently when I've autotagged a PDF, it puts all link tags inside reference tags. However, none of the links point to content within the document; they are all external web URLs or email addresses. I'm guessing that this will cause problems for assistive technologies, so we've been fixing these in the tag tree manually. My questions:

1.  Is this a real problem? Or will screen readers just ignore the reference tag and treat it as a normal external link? If that's the case we'll just leave them alone.
2.  Is there an easier workaround than fixing them in the tag tree? Sometimes we have very long documents with lots of links, so the fix will be time consuming.

Thank you!

[Bevi Chagnon - PubCom.com · Accepted solution]

Hi @Oberlin Webteam,

That's a correct tag structure for hyperlinks. <Reference> refers to an element that refers to other content somewhere else: inside the document or external.

Refer to the PDF Association's free Syntax Guide at https://pdfa.org/resource/tagged-pdf-best-practice-guide-syntax/  FYI, the PDF Association develops the standards for all kinds of PDFs, including PDF/UA for accessibility.

I'd be more concerned about the 2 nested <Span> tags that precede your sample. WTF!  Can't imagine what else autotag did to your file. Is there a reason why you're using it?

[Oberlin Webteam]

Thanks! I often use autotagging as a starting point, but as I get better at it maybe I should just start from scratch.

Before posting I checked the syntax guide, which I found in your response on another post. All of the examples of <Reference> involved internal content (TOC, footnotes, etc.), and the <Link> examples with external resources did not include a <Reference> tag. So it wasn't clear that they could be used together for external links.

[Bevi Chagnon - PubCom.com]

I can't recall any language in the PDF/UA standard itself that limits <Reference> to only internal destinations. But I'll reread it this weekend and get back to you with more definitive guidance.

[Joel Geraci]

@Bevi Chagnon - PubCom.com  - FYI - If the hyperlinks are in a different color than the surrounding text and the hyperlink has a line break, that would account for the two nested spans. One for the style change within a paragraph and one to account for the line break. Our Auto-tagger leverages our Extract API which "notices" these things.

[Bevi Chagnon - PubCom.com]

PDF/UA-1 is still a "young" standard, so the kinks and interpretations of the standard will be a work in progress for quite a while.

The committee and working groups that develop it are working hard on clarifying the tag structure for both us who make/remediate PDFs, as well as for software manufacturers and assistive technologies. So expect "refinement" as a continual process.  :-0

[Joel Geraci]

Very well put.

[Oberlin Webteam]

That's exactly what happened here – different color and a line break. Thanks for clarifying.