[Blindmath] Is the PDF format acceptable as an assistive technology

Neil Soiffer NeilS at dessci.com
Mon Jan 30 17:46:14 UTC 2012


Since Michael mentioned my name, I'll expand a bit on what he said:

First off, PDF can be made pretty much as accessible as any other format.
Most people/software don't try, so in general, accessibility is pretty bad
in PDF.  Maybe five years ago, PDF's tended to be encrypted so that not
even the text was accessible, but since Adobe changed the default setting,
that is rarely the case anymore AFAIK.  If, and that's a big if, the PDF is
properly tagged and a few other rules are followed, then there is very
little difference between PDF and HTML in terms of accessibility.  I'm on
the PDF/UA committee that has developed an ISO standard (ISO 14289) for
what it means for a PDF document to be accessible.  That's going to a vote
of ISO member countries in a month or so.  We are already working on a
version 2 of that standard that deals with math and improves table
accessibility -- that required some changes to the PDF standard itself (ISO
32000), and couldn't move forward until that version (ISO 32000-2) advanced
further in the very long ISO process.  I suspect that like WCAG, this
standard will get adopted into law in a number of countries so that
government agencies that put out PDF documents will need to conform to the
standard.  There are no guarantees on that though.

For math, I have a "proof of concept" version of MathPlayer that works with
Adobe Acrobat.  It does most everything that MathPlayer+IE does in terms of
reading math, converting to braille, etc.  It does require proper tagging
of the math with MathML, and that can only be done by hand at the moment.
One of the people working on pdftex is working on making pdftex produce
tagged PDF, and that includes tagging for math.  He's a prof in Australia
and works on the project in his spare time. It is a tough problem, but he
has made significant progress and in another year or so, might be at a
point where he can release his changes.  That would make a big source of
scientific documents produce accessible PDF.  The other two big sources are
Word and InDesign.  Getting them to change would likely require either
market pressure (possible if there's a law mandating accessible PDF) or
pressure/legal threats from groups like NFB.

All of what I've mentioned is about what may happen and none will change
the vast amount of inaccessible PDF documents that currently exists.
Looking towards the future, not counting the disappointing ibooks
announcement last week from Apple, I think EPUB 3 is where the publishing
industry is headed.  EPUB 3 is basically HTML5 with a requirement that
MathML be used for math.  EPUB 3 was designed with accessibility in mind,
so I think the future looks brighter in terms of accessible books and
textbooks.

Neil Soiffer
Senior Scientist
Design Science, Inc.
www.dessci.com
~ Makers of MathType, MathFlow, MathPlayer, MathDaisy, Equation Editor ~







On Mon, Jan 30, 2012 at 5:07 AM, Michael Whapples <mwhapples at aim.com> wrote:

> There is a lot you have said there.
>
> Firstly the main question, is PDF an acceptably accessible format? The
> answer is, potentially in theory it could be better than it normally is and
> possibly may be about as good as most other main stream alternatives. On
> the equations thing, I believe Niel Soyfer from design science may be
> involved in a project to make equations accessible in PDF, but like most
> things in this field progress is slow and I don't think anything is
> confirmed to be included anytime soon. Having said that, in practice it
> normally falls far below that and is very substandard as the author has to
> do specific things to make it accessible (many might not) and some tools
> may simply not support the accessibility stuff anyway (eg. many of the
> non-adobe PDF creation tools).
>
> <snip>



More information about the BlindMath mailing list