E-mail List Archives

Re: PDF remediation


From: Metzessible
Date: Oct 23, 2017 1:31PM

Great question! I love complaining about PDF tools.

I'd honestly never heard of PAVE, so I decided to give it a whirl. It's an
interesting concept, but leaves much to be imagined in terms of providing
an actual accessible PDF beyond what a user can do by merely using the
'Print' function on the authoring file. It's also really confusing to use,
as it will tell me there's a problem with the PDF, but it doesn't actually
show me where the problem is. I'm also irritated that error handling solely
relies on color formatting. This is probably the most basic accessibility
mistake a developer can make, which instantly made me suspect the PDF
results would be extremely subpar. All this is said with the
acknowledgement that it's hard to complain about free software, since you
get what you pay for.

I evaluated AxesPDF recently and received the following errors evaluating
the software:

1. *Syntax error:* Operator ‘h’ is not allowed in the current state.
2. *Syntax error:*Operator ‘f’ is not allowed in the current state.

Unlike free software, I CAN complain about this. For over $700 USD per
license, I'd expect it to work out of the box, and give me better error
explanations. I also expect it to be intuitive to use (The UI reminds me of
one of those old OSX X11 apps from years back), or at least provide a basic
concept of how the rudimentary workflow is supposed to be managed. It did
none of these things, and was left irritated after I tried it. I know
several other SMEs in the field use this software for remediation, but I'm
not impressed at all with the level of accessible PDF it returns with the
PDF/UA flag attached. I can go into much further detail about it if anyone
cares to hear my thoughts about it, but I won't be buying this software
because of my evaluation.

I've tried reaching out to CommonLook four times for a demo, and no one has
ever contacted me. So I can't speak about it as a solution that's any
better or worse than anything else. However, I use CommonLook Validator
(the free QA tool they provide) religiously. Quite honestly, I'll never
understand why anyone would ever rely on a vendor's included accessibility
checker to evaluate mistakes. You're only going to be looking at the things
that the vendor understands about a given problem, not according to
existing standards.

In any case, if CommonLook GA is similar to Validator, the UI might be a
bit clunky if you're already having problems training colleagues on how to
use a program like Acrobat. The errors in Validator provide no explanation
to why something is considered an error or warning, and it uses language
which requires a solid understanding of PDF that most people clearly lack
in general. Also, it cannot seem to identify pagination artifacts at all.
While it states that for the purpose of testing it marks them as artifacts
for you, it still shows up as a failure. The reports they generate are very
helpful, but this (and other issues, like missing BBox attributes which
clearly exist) must be sanitized from the final report after the fact. This
can make for a time consuming analysis, especially if you rely on these
reports for QA, Kirkpatrick training supplements, or for CMMI CM, MM or
PPQA Process Improvement evidence.

I've heard great things about Equidox. The price seems a bit high at first
glance (1200 for 10 licenses for education). However, I don't think you
need Acrobat for it, which makes it only $120 per person. I'm a bit
concerned about how astronomical the commercial license would be. The one
thing that I didn't like was that it was per-page remediation, and very
visual in how you use it. It seems to take all the things people like about
using the Touchup Reading Order Panel in Acrobat and makes it usable and
effective. Also, it's worth noting, I've never seen a PDF from this SASS,
so I can't speak to it's quality either.

The one thing I have played around with when I get chances to use it is
Tesseract for OCR. It's a bit more complex of a tool, but the results are
amazing. In fact, using Tesseract with Adobe's built-in autotag feature is
like what would happen if you found some free audio software that could
make YouTube captions with 90% accuracy. It's really powerful, yet
extremely confusing and difficult to learn.

That said, my go to tool is still Acrobat. I use both a mac and PC, but I
prefer the subtleties of the Mac Acrobat version over the PC version. When
I'm editing a bunch of tags, it's often easier and less time consuming to
single-click the tag, then click again to enter it. In later versions of
Acrobat for the PC, this seems to require extra clicks. But it's not the
end of the world, and therefore Acrobat is my go-to tool of choice
regardless. My main problem is that Adobe made the software a Jack of All
Trades, yet a Master of None. So it handles everything under the sun that
is related to PDF, but it does them equally badly. That said, one could
theoretically use the command line to remediate PDFs, but why would anyone
go through that much trouble? Acrobat is what it is, but it's still the
defacto industry standard, and will most likely be going forward for the
foreseeable future. Software companies either try to reinvent the wheel by
blatantly ignoring the ISO standards, or make something that requires you
to own Acrobat in the first place.

Though, I'm actually a service provider myself, so what works for me as a
PDF a11y expert might not work for your situation. Therefore, I have to
agree with Bevi that the best solution is to create the most accessible
authored file you can BEFORE it goes to Acrobat so you have more hair at
the end of remediation! In fact, you can automate most of your Word file
backlog using macros and VBA once you have accessible templates in place.
This makes remediation that much easier, especially since you have the
original files.

Jon Metz

On Mon, Oct 23, 2017 at 2:36 PM, Denis Boudreau < <EMAIL REMOVED> >

> Thanks Sean, that's very useful to know.
> /Denis
> --
> Denis Boudreau,
> Accessibility, user experience & inclusive design
> Cell: +1-514-730-9168
> Email: <EMAIL REMOVED> [mailto: <EMAIL REMOVED> ]
> Twitter: @dboudreau [http://www.twitter.com/dboudreau]
> On 2017-10-22 2:26:12 PM, Sean Keegan < <EMAIL REMOVED> > wrote:
> Hey Denis,
> > In your experience, how good is it at turning an inaccessible PDF into
> one we
> > would consider accessible, or at least WCAG 2.0 AA compliant? If you ran
> a PDF
> > made accessible with Equidox, would it pass the Acrobat accessibility
> > checker, or PAC2?
> Overall, it does a good job with most types of PDF documents. There have
> been one or two issues with how background images and text are recognized
> with PDFs originating from Adobe Illustrator, but these seem to have been
> outlier issues. One thing I do like about the tool is a "sensitivity"
> slider that sets a zone for the content. By moving the sensitivity slider,
> you can quickly change the regions and zones on the page to include or
> exclude text content. Another feature that works (mostly) is a Preview
> option that provides a linearized view of the page content. That helps
> identify where the logical structure of the page is incorrect.
> Yes - you can use Equidox to create PDFs that pass the Acrobat
> accessibility checker. After making corrections in Equidox, I generally run
> the PDF through the Acrobat accessibility checker to perform a quick
> verification of my edits. So far, the checker only reports the need to
> perform the manual tests. I have not tested with PAC2.
> I am encountering some issues with PDFs created from PowerPoint in that all
> images were being recognized, including those that were intended as
> background content. I suspect these background images were just dropped
> into the slides and not managed via the Master template. Their development
> team has been very responsive, so I don't expect this to be a problem going
> forward.
> Take care,
> Sean
> > > > > > > > >