A Vision for the Future of PDF

When Adobe Co-Founder John Warnock first proposed his plan for an entirely self-contained universal document format, in the Spring of 1991, he was met with skepticism. One critic is even said to have called it "the dumbest idea I've ever heard."

It wasn't unreasonable feedback at the time. Getting documents to display and print the same, in all respects and from any device had been tried many times before.

Warnock, however, felt this time was different. And in just two years, a system he whimsically called project "Camelot" evolved from cool idea to a real file format, available free of charge. In 1993, PDF was born.

A quarter-century later, it's fair to say PDF is stronger than ever. In commercial publishing, printing, document sharing, collaboration, archiving and more, PDF is the dominant final-form electronic document format.

But what will PDF look like in another 25 years, in 2043, when PDF turns 50?

The answers to these questions were on microphone earlier this year at PDF Days Europe 2018 during the summit's closing panel discussion. Catherine Andersz, CEO of PDFTron, was asked by the PDF Association to share her opinion on the future of PDF.

So what was PDFTron's take? And how is it relevant to those wondering what new direction, and what new features, PDF holds in store?

While PDF may change, the key value proposition, and the essential purpose of PDF, which propelled PDF's success so far, will likely never go out of style.

Case in point?

While competitor formats such as epub have waxed and waned in popularity, PDF has maintained stable growth, as evidenced by keyword search trends shared by PDF Association Executive Director Duff Johnson. More and more people are searching for PDF, and for ways to convert competitor formats into PDF.

Meanwhile, PDFs are still not very cool on mobile devices, with many PDFs awkward to read on small screens. Yet, paradoxically, PDF use on mobile devices keeps growing.

These trends suggest PDF's core value proposition is a matter of necessity: users will always need a way to literally "get on the same page", and thus will always need a solution for complex document sharing and collaboration. And what the numbers show is that PDF currently dominates that niche.

Making any prediction, especially one so far out, is difficult. But given what we know today, one can predict with a good degree of certainty, 25 years forward, PDF will still be used and in fact, may be used even more widely.

And PDF can only grow in importance as the digital world evolves, as cutting-edge technology becomes more accessible, as the use of mobile devices becomes more pervasive, and as going paperless becomes inevitable.

So, perhaps the real question isn't so much whether PDF's key value prop will ever go out of style, but rather how we can build on and retain the key value prop as we try to address future user requirements and push the boundaries of PDF.

It is very likely 25 years from now, PDF will have the same essential role and purpose: as an inexpensive, simple, efficient and reliable interchange of visual information.

However, "PDF has grown up. It needs new use cases. It needs to meet those use cases and the future needs of users, and what we might dream of doing next," Andersz said at PDF Days Europe. (You can watch her full presentation below.)

"There has been a lot of talk about evolving the format and adding features. But do we really need to do that? Do we really need to make the PDF format more complicated? Will it even solve the problem? That might not be the case."

While we may need to further adapt the format to accommodate future applications and users, changing PDF itself may not be the only answer -- or a true game-changer.

Since PDF's inception, document technology has made leaps and bounds. For example, compressed streams, jbig2, jpeg2000 and more, each have helped make PDF smaller, lightweight and easier to view, edit, annotate and share.

But just as how transistor-sizes can't shrink much further, leading some to pronounce Moore's law dead, PDF compression nears its theoretical limits. Further significant advances are unlikely.

Also, most new features added to the core specification since 1993, such as Tagged PDF, GeoPDF, XFA forms, 3D, JavaScript and even XMP have remained niche.

For example, anyone can create a tagged PDF. But fewer than one in six PDFs created after 2003 are tagged. And tagging the remainder poses a costly, time-consuming project that many organizations may be hesitant to take on.

Furthermore, less than a percent of PDFs currently conform with PDF/A and PDF/UA standards, respectively.

Some new features added (i.e. XFA) were even subsequently removed from the standard.

Thus, rather than extending the core PDF ISO spec -- already over 700 pages -- and waiting for the billions of PDFs in the wild to fall in line, "The next step is in helping other technologies use ... PDFs," said Andersz. And maybe, "We won't have to fix PDF -- it's pretty good already."

When J. Warnock first proposed project Camelot, he imagined a "substantially simpler, and smaller" version of the full PostScript page description language he helped design and which would become Adobe's flagship product in 1984.

IPS (Interchange PostScript) as Warnock called it, would need to be made concise and flexible to work with as many systems as possible. As he wrote in 1991:

"The right way to think about IPS is as it relates to English. No person in the world knows every English word, but a small subset of the English words, and certain usage patterns enable people to consistently communicate."

Since its inception, however, PDF has grown new features faster than barnacles. Meanwhile, PDF page description (with roots in PostScript technology) has been remarkably stable & successful. It is very possible the format would have been even more popular today had the specification been thinner and laser-focused on its core value prop.

What then, should PDF look like at 50?

Today, people are abandoning their movie and music collections for net services, like Netflix and Spotify, where the everyday user will not care to know which underlying technology or formats are being used. Users just want to enjoy their experience -- effortlessly and without limitations.

Likewise, in an increasingly format-less future, users shouldn't necessarily have to care what document format they're working with. One would expect future users may not even know they are using PDF, and that PDF will blend into web platforms and applications, becoming an unobserved part of the experience, similar to what the console window of your desktop computer is today.

In the space of a decade, PDF has the opportunity to "transcend the page" and become truly ubiquitous. And by enabling integration of PDF with emergent technologies, we can help get it to where it needs to be sooner.

The absence of semantic structure in PDF is currently one of the format's greatest strengths as well as one of its greatest perceived weaknesses, due to there being no universally accepted way to understand document structure (of which we have so many types and different complexity levels).

However, transforming PDF from being the least common denominator of file formats doesn't require changing PDF so much as the technologies around it.

Improvements in AI and related fields will undoubtedly be more of a game changer for the advancement of document understanding and re-purposing of PDF content, with significant applications in areas such as accessibility, search engine optimization, data analysis and more. Reliable reflow promises to be a key area of focus, to enable PDF text-size and layout to adapt seamlessly to whatever mobile device or browser it displays on.

Ultimately, as technology evolves, we can expect PDF to become a reliable medium of universal information exchange between different systems, and for its perceived weaknesses to fade away.

Now that the blockchain hype is fading, we can see some practical business applications coming to the fore.

A couple things to know first. While well-suited for tracking finite transactional data, blockchain faces technical limits. Limited network processing power throttles the amount of data any truly decentralized digital ledger can process at any given time.

Cryptocurrency technologists are working hard to overcome these obstacles. But it's unknown how far they can go.

In the interim, while it may not disrupt PDF, blockchain may very well add value.

Two areas are document management and verification. You can expect to see blockchain used to help protect and track PDF similar to how it is being used in supply chain management to track high-value assets.

For example, London-based company Everledger, founded in 2015, claims to have placed over two million diamonds on its blockchain, which uses IBM's blockchain tracking service to encrypt asset provenance.

With Everledger, each physical diamond gets its own digital twin, recording caret, color, and a laser-inscribed serial number. This digital copy, recorded on an immutable cloud ledger, in turn reassures buyers, who achieve a shared view of each diamond's quality and supply chain journey, without jeopardizing their privacy or confidentiality.

Likewise, Joris Schellens at PDF Days Europe proposed signing documents in the blockchain instead of storing the signature in the PDF. This would not only reduce the code complexity for developers wanting to add in digital signing or verification. The same principle could be used to streamline document workflows, keep track of a document's location, ownership and more.

In other words, blockchain could be used to help track and authenticate PDF, just as Everledger is used to reassure buyers that their purchase is not a counterfeit or a conflict stone.

Today, PDF is just a page. But PDF can transcend the page. And the technology itself does not have to change that much. As machine learning and AI make key advancements, and as other technologies such as blockchain develop, PDF will gain new opportunities to overcome its weaknesses and unlock its potential to meet the future needs of end-users and businesses alike.