WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: PDF on websites + PDF is *not* accessible

for

From: Duff Johnson
Date: Jul 9, 2013 9:51AM

Shawn,

You've called out PDF as a format for specific comment. You've made several assertions pertaining to technical and practical realities.

While I deeply respect your perspective and the conversation thus far not all statements have been accurate, informative or helpful.

This is an important forum, and I want to put the record straight.

A summary of my extended response (see below) is this:

- Shawn's confusing the PDF format itself with software implementations thereof. Contra her claim, this is the fundamental "misnomer" at hand in the discussion.

- In WCAG 2.0 terms Shawn's conflating "accessibility" with "accessibility-supported." They are rightly distinct; collapsing the distinction does not serve understanding or drive solutions to improve PDF accessibility.

- Such confusion is *naturally* understandable for end-users. Informed advocates, however, need to respect these critical distinctions to be maximally effective in terms of influencing software development - the aforementioned "implementations."

- The work Shawn's doing with TAdER is a superb contribution to critical, practical information for any developer working on user experience - including those focussed on making electronic documents accessible. The project would be more successful if it refrained from inaccurate generalizations while making technical claims.

> Background from previous comments is below [1].
>
> The problem is that PDF is currently *not sufficiently accessible* to many people with low vision, dyslexia, and related conditions and situations that impact reading - because Adobe Reader and other PDF viewers lack sufficient text customization functionality.

First, this statement is equally true for plenty of implementations of HTML / CSS / JavaScript technology, various combinations of which produce results that defeat today's AT technologies. It's hardly a "PDF problem" - calling out PDF specifically in this case is misleading.

That said, if plain HTML can meet TAdER requirements then it's hard to understand why PDF/UA doesn't as well since PDF/UA files may be readily exported to plain HTML while (key point!) retaining the ability to fallback to the real document.

The later point, however, is that you cannot reasonably claim there's any misnomer in describing properly-tagged (PDF/UA-1) files as "accessible". They are accessible within any reasonable definition that you'd apply to WCAG 2.0 conforming HTML.

But that's not my core point, which is that categorical statements such as you've offered are actually and notably injurious to the cause of promoting development in accessibility technology. You won't win new software development efforts by attacking the file-formats themselves.

Look at it this way: Accessible alternatives to PDF files will depend in the real-world on re-use of PDF content because it is the only available content in many given situations. Lots of stuff is in PDF for reasons that just make sense to the players at hand in their respective business contexts, and that's the reality with which you need to engage.

Having accepted it one ought (it seems to me) focus on highlighting the *lack* of technical barriers to full access to PDF, the degree of standardization available, the extent of implementation tools from PDFlib, iText, Adobe, Microsoft and others. Give developers reasons to invest in accessibility tools; do not foreclose on the potential for investment by proclaiming that somehow alternatives must be provided!

When tagging or re-use fails a given user for a given reason, why, the PDF is the *critical* backup, tedious as it make be, it's still the *actual* document.

That's the PDF focus and value-proposition, and it's got nothing to do with the ability to adjust kerning for X percent of users who need such. Equal access to the *same* content is - or ought to be - the focus here. There's no technical barrier to full accessibility (including TAdER parameters) to content delivered in PDF, and you do the community a disservice when you suggest otherwise.

You would achieve far more, I think, by making statements similar to the following:

"PDF files can only be considered fully accessible in every possible use-case when full TAdER text management is available."

Only when you are using such constructive terms will you will then be helping, rather that hurting, the cause of improving access to content that exists (fact) in PDF.

> Even well tagged PDF that is more accessible to screen reader users is still *not accessible* to many people with other print disabilities. Accessibility is more than screen reader access.

This is a straw-man. No-one claims this for PDF any more than for HTML.

> Unfortunately, "tagged PDF" started getting called "accessible PDF" -- that is inaccurate and a harmful misnomer.

Let's be completely clear on this.

PDF that conforms with PDF/UA-1 (ISO 14289-1) is accessible, period. Whether it is accessibility-supported for any given need is *another matter* - a question for implementers, but assuredly not in question with respect to the file format.

If vanilla HTML meets TAdER or TAdER-relevent standards for accessibility, then it's fair to say that PDF can qualify, because PDF/UA-1 conforming PDF files can be exported to the user's chosen HTML implementation. As you rightly point out, forms aren't yet supported in VIP Reader. Why is this more worthy of note that X, Y or Z failure of FaceBook, or iTunes or whatever to accommodate TAdER preferences?

Do you really want to suggest that PDF is inaccessible now but can "become accessible" when (for example (VIP Reader adds the ability to print? C'mon.

Now, anyone can and will concede that it's also possible to simply "tag" PDF without doing it well, just as its possible to do a half-ass job of ensuring CSS usage is accessible (for example). Tagging is a critical but insufficient part of meeting accessibility requirements for PDF content. Tell them that, instead of trying to make an (incorrect) grossly general claim about the file-format as a whole.

> It perpetuates the lack of awareness, even among accessibility specialists, that PDF is actually not accessible to many people with print disabilities.
>
>> My job is to communicate one person's ideas to another person.
>> I want to provide what is both legally required and what is desirable to the users.
>
> While PDF is a useful medium for some situations; when it is used, there must be a more accessible alternative provided in order for the information to be available to people with disabilities.

Once upon a time this was true. Today, this claim is inaccurate as written. It would be much more useful if it were re-written as follows (for example):

"While PDF is a vital medium for many situations; when it is used, PDF creation, editing or reading software should create or read PDF/UA in order to generate or provide appropriate alternative representations of the PDF's content in order for that PDF's actual and inherent content to be available to the greatest possible number of people with disabilities."

> I've been fairly quiet about this for many years (except to Adobe product managers :) because the accessibility of PDF has improved from years ago, but I'm deeply concerned about the *misconception that PDF is accessible*.

PDF *can* be made accessible, and at this level of generalization, it's absurd to highlight PDF in format terms. Plenty of large-scale HTML implementations don't meet your preferred specifications for full accessibility today but I don't see you calling out HTML per se as "inaccessible."

> For more info, please see:
> * Text Customization for Readability <http://www.tader.info/>;

I commend you for this site, and would unreservedly recommend that reader software implementers - regardless of medium - review your substantive recommendations for UI controls in readers.

> * PDF viewers section of Support for Text Customization <http://www.tader.info/support.html#PDFisNOTaccessible>;
>
> (That is a work in progress and I welcome feedback directly.)

Well, I've been no less public that you, I guess, with my opinion! ;)

Duff.