Acrobat: The disparity of tagging methods

Prompted both by troubleshooting a comrade’s accessibility work (related to this short RPG collection which you should absolutely check out!) and a recent instance of tags in a work document turning to random bytes, I thought it might be valuable to briefly go over the three main ways to tag elements in a tagged PDF in Adobe Acrobat. Ultimately they should all do the same thing, but because it’s an Adobe product, they all come with their own unique quirks.

It’s worth noting that tags are one of a handful of structural components in a PDF. While they are probably the most important for accessibility, they need to play nicely with the structure at play at the Content level. While slightly irrelevant to this discussion, it’s also worth a reminder that tagged PDF in PDF 1.71 is described in §10.7, pg. 841; standard tags described in §10.7.3, pg. 898.

Simply retagging things in the Tags panel

This is the safest bet, but it assumes that a: the document is already tagged (just incorrectly), and b: the items that need to be retagged are already grouped properly at the container level. If the document is untagged, and no source is available from which to generate a(n improperly, even) tagged document, then the Accessibility menu has an option to ‘Add Tags to Document’. I would only do this on a backup copy, and would give the document a good once-over to make sure fixing its decisions isn’t more work than starting from scratch with one of the two options below. On short documents, starting from scratch is often quicker, I find. On longer or more complicated ones, not so much.

At any rate, changing tags in the Tags panel is generally quite quick, especially with a working knowledge of what the standard tags actually are. There aren’t a ton of obvious keyboard shortcuts, but a while back I wrote about important ones that help me speed through the process. While items that need to be broken up at the container level need a different approach, it is worth keeping in mind that often several containers will have ended up in one tag, and these can freely be moved about.

The key advantage to this method, aside from speed, is that it does not touch things as far as the Content panel is concerned. This is very important; it means you’re only fixing stuff in one panel instead of hopping back and forth between two (which can generally be achieved quickly with the respective panel’s ‘Show in [other] Panel’ option, but still).

The Touch Up Reading Order tool

A graphical approach wherein one rectangularly wraps objects and tags them via dialog box or contextual menu, this tool has the added benefit of theoretically reordering the reading order of things. Reading order in PDF is complicated, and based on Content order and Tag order. The Touch Up Reading Order tool affects the Content order, which means it moves stuff around on the Z plane. If you have background items like solid blocks of color, using the Touch Up Reading Order tool on text above it is likely to hide it underneath. Then you need to reorder these elements manually in the Content panel.

Other than that, it generally works well. If you’re dealing with straightforward text that doesn’t overlap any graphical elements, you’re unlikely to run into issues with this tool. Even if I’m not using it, I generally leave this mode selected so I can get an overview of what’s tagged as what on the page itself.

Using this tool on an untagged document was what caused the aforementioned random bytes issue. The document came from Publisher, most of the bytes were non-printing. I suspect it was an encoding error. Even though tagging the document this way would have been quicker, it was unusable because of this issue. Generating tags with the ‘Add Tags to Document’ tool did not cause this problem, and I was even able to use the Touch Up Reading Order tool to fix some table issues after everything already had a base tag created. I don’t know how common this problem is (I’ve only experienced it a few times and only with documents from Publisher), but I think it’s worth noting.

Select text in document, choose ‘Create Tag from Selection’ from dropdown menu in Tags panel

This is kind of my last resort if neither of the other strategies work, or if using the Touch Up Reading Order tool would cause too much additional reordering in the Content panel. Aside from being a bit slow and clunky, this tends to cause a problem wherein it starts duplicating and nesting containers in the Content panel. Despite the tags (in the Tags panel) being correct, this still seems to cause issues when interfacing with screen readers, so generally some manual reordering of things is necessary in the Content panel. What it doesn’t seem to get wrong is reordering objects in the same way that the Touch Up Reading Order tool does, tossing text behind graphics and the like.

  1. PDF 1.7 was the first release as an ISO standard. Adobe keeps doing their own shit, and now the PDF Association has kept the motion going with PDF 2. Adobe is still working off of PDF 1.7, adding proprietary garbage onto this standard, presumably because opening the format up was a complete farce to them. PDF 2, while a good standard, faces difficulty with adoption, by my observations. Despite being an ISO standard, PDF 1.7 is available for free PDF, and I implore everyone working with the format to read the key bits of this document. ↩︎