Re: [EXTERNAL] Re: SC 1.4.10 Reflow and PDFs

Thanks for that reply Duff. Very useful information and clearly presented.
________________________________
From: Duff Johnson <duff.johnson@pdfa.org>
Sent: Friday 1 December 2023 14:24
To: Mark Magennis <Mark.Magennis@skillsoft.com>
Cc: w3c WAI List <w3c-wai-ig@w3.org>
Subject: [EXTERNAL] Re: SC 1.4.10 Reflow and PDFs

You don't often get email from duff.johnson@pdfa.org. Learn why this is important<https://aka.ms/LearnAboutSenderIdentification>
A PDF doesn’t reflow in a web browser

That’s a choice made by browser implementers, not a fact about PDF.

The difference between what the PDF provides and the capabilities of viewing software is important to bear in mind. After all, once upon a time browsers didn’t support PDF at all; today all browsers include some degree of native support for PDF.

So the real question is: when will browsers get around to more advanced support for PDF?

(I’m making this point to provide background to your question, not making any claim regarding current PDF software in the 1.4.10 context. I’m not making an “accessibility supported” claim here…)

But PDF reader applications like Adobe Reader have reflow functions that reflow the text both when you zoom and when you narrow the window. However, the Reflow function in Adobe Reader removes all the images in PDFs I've tested. Which means there is loss of information, which fails 1.4.10 reflow.

This is NOT an endorsement… but worth noting that Adobe produces software which reflows in several ways… and, as with browsers, these are simply implementation choices, not facts about PDF.

And so...

- Adobe “original” reflow feature does just as you say; no support for images in reflow
- Adobe's “liquid mode” reflow feature does include images (albeit at this time it does not respect the PDF’s tags :-(

It should be noted that a specification for reuse of tagged PDF as HTML was published by the PDF Association some time ago, and is presently under review for an update:

https://pdfa.org/resource/deriving-html-from-pdf/

Software (e.g. browsers, to use your example) implementing this specification would, of course, be able to support any sort of reflow needs, images, text-spacing… whatever.

Accordingly I encourage those interested in the subject to address this problem by encouraging browser (and other) developers to take better advantage of the PDF format’s existing capabilities.

So it seems like there may not be a way for users to reflow PDFs at all without losing info. Is that right, or is it maybe to do with how the images in the PDFs I've looked at are encoded?

There are ways; the issue is implementation. The manner in which images in PDF are encoded is generally not the limiting factor.

Of course, it’s also possible to create PDF files in which, for example, individual images are not discrete objects on the page but are instead located on a larger background canvass… this sort of PDF would not be reflowable per-se, and thus would result in a “hard fail” of 1.4.10 irrespective of software. As always, it’s possible to create inaccessible content; PDF (for format) can’t prevent an author from doing so any more than can HTML.

Is it possible to include images in a way that they don’t disappear when you use Adobe Reflow?

It’s not about how the images are included in the PDF; it’s about the capabilities of the PDF viewer.

Or are there other PDF reading tools that reflow text without removing images?

As mentioned above, Adobe’s Liquid Mode (on their mobile reader) is one such. Other apps that support more advanced reflow of PDF are produced by Foxit, WPS, Moon.

The more users ask software developers for this functionality the more it will become available.

Duff Johnson
PDF Association
pdfa.org