How can I make a single PDF with internal hyperlinks, standard PDF tags, and reflow compatibility?

[Andrew Welker]

(This post is related to but different from another question I'm asking as a separate post. This is intentional.)

I'm in a bit of a pickle. I'm writing a big book of stories, adventure prompts, and optional rules for a popular board game. I'm publishing the untagged PDF and print versions. I'd like to offer a "flexible format" style PDF too, which would be a PDF file format that meets the following criteria:

• It's mobile friendly. The page is the size of a mobile phone screen instead of a normal sheet of paper. The font size is still legible! It's just got a lot more pages now than the print version does.
• You can navigate the PDF easily without using bookmarks, relying solely on internal hyperlinks (to text anchors) and buttons (that On Release or Tap will run a "Go to Previous Page" or "Go to Previous View" action). Some of these are in parent pages as a de facto footer, some are inline with meaningful content.
• It's assistive technology friendly. The PDF has a logical tag structure using the PDF 1.7 file type nomenclature, like <P>, <H5>, <Reference>, <Lbl>, or <BlockQuote>, making it compatible with popular "screen reader" text-to-speech computer programs.
• It can reflow in Adobe Acrobat (Ctrl + 4, or Menu > View > Zoom > Reflow).
• I can set up the tags and other technical details in InDesign, with little to no correction post-export in Adobe Acrobat Pro, NVDA, Microsoft OneNote, Amazon Kindle, or other document reading programs later. (Note: I'm fine with testing prototypes of this file in programs like these, I just want that once I've done the process right, I can press the "Export" button in InDesign and it's done or nearly done).

My thinking is that if I can compile the "mobile friendly format" and "reflowable and accessibly tagged format" into one file, that cuts down the number of files readers need to sort through and (more importantly) gets accessible tech into the hands of people who didn't know they could benefit from that kind of thing. Like me! I never thought I needed a screen reader software, but then in college I discovered ones built for sighted users with learning disabilities like ADHD or Dyslexia. I'm thinking of things like NaturalReaders.com, Kurzweil 3000, Microsoft Immersive Reader, Bookshare, and more. They've absolutely changed my life for the better, I owe my college education to them for getting me through dozens of chunky textbooks. My hope is that by combining the two file formats together, readers who already know they need assistive tech formats will be able to find and use it easily, people who need the mobile-friendly version and wouldn't benefit much from assistive tech can find and use it easily, and people who don't know they'd benefit from assistive tech perks but would like to try the mobile-friendly version can have a much easier time stumbling upon such helpful features.

I've a lot of research trying to figure out how to do each of these bullet points alone in a PDF, but I'm having a trickier time doing them all at once. I think I can make it mobile friendly and tagged either way, but "internal navigation through hyperlinks and buttons" and "reflowable in Acrobat" seem to be mutually exclusive. Here's my ideas for potential solutions so far. Does anyone have any further ideas, insight, or suggestions?

• Give up on having a PDF that can be navigated through in-document links (not bookmarks) AND that can reflow in Acrobat, and make two separate documents: a mobile-friendly PDF with PDF tags and links, and a reflowable EPUB using the same PDF tags but no links.
• Find a way so that when the PDF goes into Reflow mode, it effectively "turns off" interactive annotations in the whole document, displaying the relevant text links and buttons identically to how it normally would, except the links don't work while it's in Reflow mode. So inline "see chapter 4" kinds of internal text anchor hyperlink references turn off although the text remains, and parent page artifacts would be missing entirely. But only during reflow mode.
• Find a way so that links still work without the "interactive annotation" status in the PDF's programming, and either a way to generate all the links at once without generating this status or a way to delete the status for all the links all at once post-generation. This may mean they still work in Reflow mode.

[Eugene Tyson]

Quite a lot to understand here - I've been reading it and rereading it so let me try to help here

Creating a single PDF that's perfectly mobile-friendly, internally hyperlinked, fully accessible, AND reflowable in Acrobat is extremely challenging, and likely involves compromises.

PDF reflow mode and interactive hyperlinks/buttons don't play nicely together. Reflow disrupts the fixed layout that hyperlinks rely on.

I think you may need two separate documents or at least layouts @Barb Binder explained this method beautifully recently, and might chime in, I can't find the thread now. 

I think you need a Mobile-Friendly PDF (with Links & Tags) focussed on readability and in-document navigation using hyperlinks. Tag it well for accessibility etc.

Then you might need a  Reflowable EPUB (with Tags, Simple/No Links): EPUB is designed for reflow. Navigation will be simpler (Table of Contents), and complex hyperlinks/buttons might not work well. Best for screen reader users and reflow.

You may have to manage two files or at least multiple layouts .

@James Gifford—NitroPress is a whiz with epub and might have some ideas for for you (and me :D) 

Single "Compromise" PDF won't be perfect. 

You'll need a

• Mobile-friendly page size.
• Internal text hyperlinks (minimize buttons/complex navigation).
• Full accessibility tagging in InDesign (crucial!).
• Accept that reflow might be imperfect; hyperlinks may be visually disrupted or less functional in reflow mode.

I don't know the best user experience across the board, maybe creating two separate documents (PDF and EPUB) is likely the most practical and robust solution. If you must have a single PDF, be prepared to make compromises and testing.

Looking forward to reading more about this conundrum!

[Andrew Welker]

You, good sir Eugene, are a godsend.

My understanding is that buttons (at least when they're made out of a text frame with text in it, as opposed to a graphic frame or other element) and hyperlinks that go to text anchors hypothetically should work just fine in reflow because they lead to text, not pages or programming. The buttons I was trying to use, however, led to programming (specifically an action to go back to the last viewed page) and not text anchors.

It sounds like you and I have reached the same conclusion. It seems (at least theoretically) impossible (or at least prohibitively difficult to figure out) to have both reflow AND these buttons throughout one document. But reflow and internal text anchor hyperlinks seem to work together just fine! So yes, one PDF that's reflow compatible, tagged properly, has otherwise good internal navigation links, and has a small page size.

Here's the trouble I'm still having right now: I think I now can tell where to tag things in the PDF as P, Hn, and Artifact, and I think I can tell how to add in custom InDesign Structure/XML tags that get mapped to those few PDF tags upon export but can be quickly and easily remapped to more obscure PDF tags in Acrobat (eg L or BlockQuote). I'm just not sure how to ensure that all the different nesting requirements of the more obscure tags are quickly and accurately exported properly!

Example A. I know that an entire bulleted list would be tagged with an L container, that each entry would be its own LI container tag (list item), and that the bullet characters within each entry would have the Lbl tag (label). I'm not sure if the entry's proper text itself should be tagged as P (nested within LI, nested within L) or left untagged and just kind of assumed to be part of LI (nested within L). If the list has a name of some sort, I'm also not sure whether I should tag that title as a context-appropriate Hn (occurring just before L) or a Caption (nested within L).

Examble B. I've got similar concerns for inline hyperlinks that are internal references. If I have a part of my board game rulebook that say "the enemy uses the statistics of a mobster (see chapter 4)", I know that the boldface visually indicates that it's a reference to a specific game mechanic you can find described in more detail in another part of the book (chapter 4). So if the paragraph it's in as a whole is tagged as a P container and I put a hyperlink on "mobster" or "mobster (see chapter 4)", should that text be tagged as: (a Link within a Reference within a P) or (a Reference within a Link within a P)? Or should "mobster" be a Reference while "(see chapter 4)" is a BibEntry, and both are nested within one Link tag, or only one has a Link nested within, or each have separate Links nested within going to the same place? Are any of these possible combinations definitely wrong or definitely right, or are they all about as equally effective?

I'm also planning on releasing a reflowable EPUB as well, yes. Regrettably, the tags for that file format are similar but subtly different from the ones for PDF and I'm not yet sure where to go learn how and in what ways. Or how I can adjust them after exporting from InDesign as needed. But thankfully I can test them on mobile via my iphone's Kindle app and on desktop via that Kindle app and Adobe Digital Editions, and I think I know where to edit them pre-export in InDesign. 

Currently my plan is to offer the following formats: standard PDF, "quick load" PDF (black and white, little to no decorative graphics), "flexible format" PDF (this mobile + reflow + tagged friendly one), and reflowable EPUB (it's just an EPUB of the last one's files).

I've been finding a lot of promising guidance from things published by the PDF Association (pdfa.org). Their Techniques for Accessible PDF (UA-1), Matterhorn Protocol, and Tagged PDF Best Practice Guide seem to be very promising. Also a few of their conference presentations uploaded to their youtube channel have been helpful, like "Accessible PDF - How to tag content the right way (Klaas Posselt)" has the first visual representation of how to nest tags properly that I've seen. That diagram is what suddenly made things click in my head to realize what I'm still missing!