Duff Johnson’s PDF Perspective

Organized by Christopher Smith’s AGI (recently acquired by Aquent), last week’s 2007 Adobe Acrobat PDF Conference in Orlando, Florida was a notable success.  The co-location with the CRE8 conference for graphic designers expanded the scope to the benefit of all attendees.

The Scene

From the opening welcome party complete with Mickey Mouse and open bar, to Al Gore, Marissa Mayer and the mid-conference networking party (also well lubricated) to many fine sessions including the Acrobat Alternatives melee and rowdy Power Panel, this conference was both educational and highly enjoyable.  Adobe offered a very tasty (if very early!) Thursday breakfast for members of Adobe’s Acrobat User Groups.

The Keynotes

While the PDF Conference may be one of the smaller platforms for Al Gore’s famous presentation on global climate change, the former US Vice President gave it his all. He delivered a convincing demonstration of both the facts of our planetary predicament and his passion in communicating on the subject.  More than a few attendees noted a substantial improvement in the former Veep’s silhouette.  Having cut a Hitchcockian figure at the recent Academy Awards, Gore has clearly been working the problem in the gym.  Let’s just say he was inspiring in more ways than one!

Regrettably, Mr. Gore didn’t spend much time discussing PDF (his preference for presentations is Apple’s Keynote), but he did take note of FormRouter’s GreenPDF initiative.  Don’t print it; PDF it instead, and help control global climate change!

Google’s Marissa Mayer, Vice President for Search Products & User Experience, enraptured the crowd with a simple but telling survey of the company’s activities. Her presentation described how Google’s goals are organized around the concept of “responsiveness”, which Google measures in microseconds and considers a (if not the) key metric in almost every application.  Just how they get their servers to respond faster than my tricked-out Windoze machine can do locally, I suspect I’ll never know.  I’m just glad I bought the stock.

From Section 508 to copy-and-paste

Nettie Hartsock of Planet PDF managed to endure my own session without (it seems) terminal boredom.  It’s not easy to make PDF accessibility a scintillating subject, but I had an attentive audience, one several times larger than last year’s talk on the same subject.

The benefits of accessible PDF extend well beyond disabled users, and folks appear to be catching on, even outside Washington.  There was a great deal of interest in how PDF accessibility affects search-engine performance, and a lot of nodding heads when I described how and why PDF files are characteristically ignored by web content managers.  I’ve posted my presentation (pdf, 555 kb) for your reading pleasure.

Other News

Adobe has posted a Vista FAQ answering (in part) the growing chorus of users wondering when Acrobat 8 will work properly under Vista.  I may now quote Adobe as stating that: “In the first half of 2007, we [Adobe] expect to issue a free update to Acrobat 8 to support Vista.”

PDF was originally designed to do one thing: deliver an author’s intent to screen or printer in an efficient, precise and platform-independent manner. This “print paradigm” persists today, and colors the issue of accessible PDF in subtle but pervasive ways.

To understand how the print orientation in PDF presents a challenge to accessibility, begin with the fact that a PDF has no intrinsic notion of words or paragraphs.

PDF “thinks” like a printer “thinks”. Objects (such as text, images, lines, etc.) appearing on the page are considered in terms of precise location and sequence of appearance, not their semantic relationship to one another. Words and paragraphs aren’t useful constructs in printing, for they offer no assistance in the placement of ink.  As a consequence, the very concept of a word was literally missing from PDF until 2000, when Adobe published the basic mechanism for “structured” PDF with the PDF Reference 1.3.

What is structured PDF?  Simply put, it is PDF with additional information to organize the objects on the page into words, lines and paragraphs, and to order these larger blocks of content into a “text flow”.

In late 2001, Adobe published the PDF Reference 1.4, which described ”tagged” pdf, adding organization and nomenclature to the structure elements.

Tags allow the author to manage text-flow, define headings, add alternate text to images, ensure tables match the intent, and generally ensure that the document contents are fully and discreetly available to users who can benefit from such information.

Who are these comsumers of of structured and tagged PDF?  They include disabled and other AT (assistive technology) users, but anyone trying to copy-and-paste text or read a PDF on a mobile device will prefer tagged to untagged PDF.

Beyond strictly human “users”, tagged PDF improves search engine results, and even makes possible the extension of semantic Web concepts to PDF.

Those interested in more specific or technical information should check out Mark Gavin’s presentation, PDF and Accessibility (PDF, 857 kb), at labs.appligent.com.

In a forthcoming post, I’ll lay out the business case for better tools to enhance PDF accessibility.

The PDF/Universal Accessibility (PDF/UA) Committee met at the AIIM Conference and Expo in Boston this past week.  If you are interested, you can read the agenda from the public portion of PDF/UA’s Wiki.

We welcomed new members and had a productive day-and-a-half of meetings covering a wide range of issues, from MathML to table structure, multimedia and programmatic validation.

For those who don’t know about PDF/UA, we are a Standards Committee of AIIM, tasked with the mission of developing a draft ISO specification defining what it means to create and validate a PDF as “accessible”.

Accessibility isn’t just about screen-readers for blind and other disabled users. Accessible PDF is more usable, more navigable, friendlier to search engines and reflows smoothly onto mobile screens.  If you or your organization has an interest in the multifaceted opportunities for enhancing end-user experience of electronic content, I invite you to come join PDF/UA for one of our regular teleconference meetings!

I hear some version of the following question over and over:

“Which software accurately converts PDF to Word?”

Converting PDF to Word (or other word-processing applications, HTML or whatever) is not a simple, push-button affair, as almost everyone who has ever tried it knows (thus the questions).

Even so, most people are looking for a simple, push-button way to get the contents of a PDF into a Word file.  What’s the typical experience? Documents with layouts even slightly more complex than vanilla paragraphs routinely convert into junk. End-users expect this task to be pretty easy - which explains why the tone of the typical inquiry may be characterized as “pained”.

Let\’s take a moment to understand why converting PDF to Word is so problematic.

The factors influencing the quality of conversion from PDF to Word are, in descending order of significance:

1. The extent to which the document\’s logical structures are represented within the PDF (tagging)
2. The complexity of the objects on the page (mathematics, charts, graphs, etc)
3. The complexity of the document layout

Factor 1 is a property of the PDF file itself, not the software used to extract the contents to Word.  If the document is properly structured and tagged, predictable results may be had in converting to Word from Adobe Acrobat.

Beyond Factor 1, different software will guess at logical structure via analysis and assessment of the layout, fonts and objects on the page.  There is no magic bullet. The more complex the document, the lower the chance of high-quality output, no matter what software is used.

When I first tell someone that blind people can read PDF files, I often get a slightly puzzled expression. Sighted people sometimes appear to assume that the blind can’t read; that maybe somehow blind professors and programmers and lawyers are born, not taught.

I digress. The fact is that blind and other disabled individuals can not only read PDF files, but Word files, web pages and other electronic content using “screen reader”, or other types of “assistive” software or hardware, depending on their disability.

Beyond the simplest of documents, what is required in order to be considered “accessible”, document contents must be structured such that tables, images, footnotes and so on are correctly identified to the user. Without structure, documents are just a heap of words - or letters, if you prefer to assume away the structure that binds the letters and words together into paragraphs.

Sighted users get to “cheat” by gaining clues about tables, lists and so on from the page layout. Those who must use alternative reading methods to read the same document often can’t access this presentational information. They are dependant on either the structure (or lack thereof) in the document, or else the capacity of their reading software to correctly impute the structure based on a programmatic examination of the page.

Section 508 requires that some minimal amount of structure information must be present in order for a document to be considered compliant with the regulation.

Although PDF content is a staple on virtually all government and corporate websites, actual Section 508 compliance assessment and subsequent correction of PDF content remains poor in federal and state government, to say nothing of the corporate world. There are three principal reasons for this.

1. Generally speaking, authors of PDF documents (who aren’t web-content managers) have no idea whatsoever about ensuring their document is accessible. Web content managers, on the other hand, usually know something about accessibility, although they generally focus only HTML.
2. With existing technology, PDF content is far less amenable to automated or even semi-automated accessibility evaluation and correction as compared to content that’s delivered as tagged text (HTML, XML… all the “MLs”). There are a variety of reasons for this, but trust me, it’s true.
3. While PDFs are unambiguously “web-content”, web-content managers nonetheless tend to disregard PDFs, simply because “all they ever do” with a PDF is link to it, or facilitate the creation or emailing of a PDF. Regarding the contents of the PDF and the accessibility thereof, they generally don’t have a clue. PDF is thus the “blind spot” for web-content managers.

Without a doubt, the tools for evaluating and correcting PDF tagging to ensure Section 508 compliance need more work. Strong third-party options for PDF tagging such as NetCentric’s CommonLook Acrobat plugin are emerging.  Adobe’s Acrobat Professional 8.0 itself got a rather modest upgrade in the accessibility department - more on that in some other post. ABBYY’s FineReader and Nuance’s OmniPage OCR software are providing new tagging options for PDFs that begin life in a scanner.

The real problem at this juncture is not the developers - after all, they respond to the demand, and demand is coming, thanks in part to the new Target lawsuit. For the moment, we can safely note that it is the web-content managers who have yet to take on board what accessibility standards really mean to them, whether Section 508 or any other mandate. Section 508 doesn’t exactly set a high bar, being somewhere south of WCAG 1.0.

Reports, documentation, manuals, presentations… content needs to be tagged with semantic structure to comply with Section 508. It’s not just HTML and web pages, this applies to Word, Excel, Powerpoint… and of course, to PDF as well. Content tagging should go well beyond current Section 508 standards if a document is to be considered authentically accessible to those who must use assistive technology to read. End-users MUST learn how to structure documents correctly from the outset.

Launching the revolution in structured content is the next frontier. Let’s get on it!

Interested in accessibility for electronic content like PDF, Flash, movies, podcasts, etc?  If you are in the publishing or government sectors, you should be.

Joe Clark’s new Open and Closed Project plans to write a set of standards for the four fields of accessible media – captioning, audio description, subtitling, and dubbing. 

Who is Joe Clark?  Author, consultant, captioning guru, WCAG gadfly, active PDF/UA member and generally all-around intellectual warhead, Joe is exactly the kind of smart and fastidious purist who just might do a GREAT job doing exactly what he says he’s going to do.

In his own inimitable and irreverent way, Joe needs your help, or more specifically, your patronage (as he puts it).  He’s set up the Joe Clark MicroPatronage Project, where you can help fund Joe’s work on the Open and Closed Project.

As chair of PDF/UA, an AIIM Standards Committee in which Joe has participated for two years, I can personally attest to Joe’s… uh… forthrightness, dedication to purpose and capacity to contribute.  He’s both intellectually vigorous and rigorous, and his research is very likely to be not only interesting, but of real value to those in the field.

So, go ahead!  Make a deposit in the Karma Bank, and let’s see what Joe comes up with!

I’m often asked “Hey, when is the PDF/UA Committee going to tell us what exactly constitutes accessible PDF is and how we get there?” I don’t have an answer for this question, yet. I can tell you that we are working on timelines now - stay tuned.

In the interim, I’d like to offer, with the kind permission of the author, Adobe Systems’ Greg Pisocky, his guide to Creating Accessible PDFs using Adobe Acrobat 7.0 Professional, now hosted on the Document Solutions, Inc. website. (Full Disclosure)

IMHO, Adobe themselves should host this document at their Accessibility Resource Center, because it’s the best Adobe publication on the subject to-date. Until they do, however, we’ll just try to help out a little.

Two great minds, etc… As I was posting on the subject of PDF/UA (near and dear to my own heart), Leonard Rosenthal had just written a summary of the PDF Standards currently under development.

AIIM’s PDF/UA Committee needs your help!

PDF/UA (Universal Accessibility) is tasked with developing a Standard for accessible PDF with a goal of eventual adoption by the ISO.

The Committee is made up of volunteers (don’t I know it!), and members may be of Observer or Voting status.  Voting status requires participation.

The Committee’s Wiki is available to members of the Committee, but the meeting agendas and minutes are publicly available. What we need in particular is more involvement from both the PDF and the assistive-technology developer communities - the people who don’t mind plowing through the Reference.

If you are…

a)  Are interested in electronic content accessibility, and (in particular) accessible PDF, and
b)  Are versed in the PDF Reference, or
c)  At least know what the PDF Reference is, and
d)  Would be willing to take some time to read and discuss portions thereof

Then please join the PDF/UA Committee!

At this point, our process involves going through the PDF Reference, identifying areas of significance/concern, and then developing the Accessibility perspective on each of these, along with (suggested) revisions to the Reference.  Come help us develop the standard for accessible PDF!