Skip to content

History of PDF

Does anyone remember the concept known as paperless office? More than 25 years ago, to be concerned about this idea meant to be at the cutting-edge of technology. By the way, there is a (quite well) known effect called hype cycle that shows what happens after each technology break-thru: shortly after announcement, expectation for the consequences of the new invention grow exponentially, until they go flat, fall down as quickly as they rose, and slowly begin growing again, but this time driven by a new “genuine”, more realistic expectation. PDF invention does not escape this phenomenon, as it was presented just in the peak of the expectations generated by the paradigm of the “electronic document”, and it also had its ups and downs. I want to share a small research on this core technology of graphic industry.

The problem

At the beginning of 1990, with PostScript already consolidated in the industry, John Warnock began worrying for an ever-growing problem inside big companies’ IT departments: How can we confidently exchange documents in electronic form? Questions like this may seem trivial today, but ask an employee of such a company how to make a certain document created in a DOS or Windows environment, for instance, not only show in another, different system (let’s say a Macintosh workstation or a UNIX system) but also look identical to the original as well…

John Warnock. Imagen bajo licencia Creative Commons Attribution-Share Alike 3.0 Unported.
John Warnock, one of Adobe founders and mastermind behind PostScript and PDF.

Solution was then straightforward: printing the document. This way it had to be recognized, at least in this context, paper’s superiority over its electronic counterpart. It is not a coincidence that PostScript had reached an remarkable evolution, as it helped to improve the quality of electronically-generated printed documents.

But, why stopping in the middle? Why having a huge development in electronic document generation if it would end up on paper? Why not imagine an electronic equivalent to paper?

Today we are aware, for several reasons, that we need to reduce our paper consumption, but at that time it was kind of an avant-garde idea. While today we are still wondering all the consequences of the revolution of social networks and the Web 2.0, at those times changes that proliferation of personal computers would bring were heavily discussed (“one in every office” someone had prophesied, similarly to Henry Ford’s “a car in every garage” at the beginning of this century). One of these changes was the paperless office, predicted in a 1975 Business Week’s article. More than 40 years after being announced, the latter wasn’t fulfilled: mankind had never seen such an explosive growth of paper demand before.

Let’s turn back to John Warnock and his worries. How do we solve the electronic exchange of documents? It would be necessary to conceive a way of keeping their contents in digital form, and a way of presenting them on screen when needed. Warnock had a solution at hand for the former: PostScript had been created for this purpose, since it could describe anything a page contained. On the other hand, to actually “see” the content of some document, it had to be printed. However, by that time Adobe had developed Illustrator as a “visual PostScript interpreter”, and it was a program running both in Windows and Mac; so it was already technologically feasible to have a digital format that could be visualized, without changes, in more than one computer platform. With these ideas, Adobe started the development of a solution:

“We have to overcome our current inability of exchanging information among machines, systems and users in such a way we make sure our files look always the same no matter where they go.” (J. Warnock) [1]

In fact, Adobe had already extended PostScript to process documents intended to be seen on screen instead of printed, a technology called Display PostScript, but required computational resources were no different from a printer interpreter, which proved not to be a practical solution for the scarce power of an office computer.

The solution?

The “Camelot project”, an analysis that Warnock presented to Adobe on what he believed there were the problems with documents processing, along with an idea of how to solve it, was a first step. In that report he proposed to turn PostScript into a “lighter” version, not requiring a full interpreter to display a document on screen. Warnock introduced it with a proposed name Interchange PostScript or IPS. After explaining how it would work, Warnock said:

“Imagine being able to send full text and graphics documents (newspapers, magazine articles, technical manuals etc.) over electronic mail distribution networks. These documents could be viewed on any machine and any selected document could be printed locally. This capability would truly change the way information is managed. “ (J. Warnock, The Camelot Project) [2]

Warnock Camelot_Page_1
The Camelot project.

Realizing the need to meet some potential client to show the idea, Warnock hired Bob Wulff, a programmer with a high expertise in graphic computing and Windows applications (something unusual inside Adobe, plenty of Mac programmers). It’s been told that Warnock literally bumped into Wulff in the middle of a hallway, and right there Warnock asked him to join the team to develop a demo software to show to IBM. (In fact, Warnock asked Wulff to join them for just “a couple of weeks”; after that Wulff leaded the team that brought Photoshop to Windows. Today he is a senior engineer at Adobe.)

Adobe didn’t hide this development; pretty much the opposite, the IPS concept was publicly announced for the first time in the Seybold 1991 conference. At that year’s fall, when everything was ready, they set up a company-wide demo in the Mountain View building B lunchroom, “back when the entire company could fit in the lunchroom“, according to Wulff. “We demoed viewing files. We even gave the entire company a T-shirt. This project was for real.” [3]

Exchange1.0
Splash screen of the first PDF editing enabled software, Acrobat Exchange.

That demo turned to be a huge success. Development of a commercial version started first under the code name Camelot, after Warnock’s report, although it was later renamed to Carousel. (There is evidence of that even today: those old Mac users like me might remember that the four-letter file type code of a PDF document is “CARO”). Documents created with Carousel were named as Portable Document Formats (PDF for short), and it first version was officially announced at 1992 COMDEX expo, being awarded as “Best of COMDEX”.

After that first version was ready, the product name had to be changed: Carousel was already registered by Kodak. It was decided then to change that name to Acrobat, with the idea of something that combines skill and strength, attributes perceived by the development team to have, and it was released to market on June 15th, 1993, along with an 8-page ad in the Wall Street Journal. (Just for curiosity, Adobe published the PDF Reference book as a printed one.) To be more specific, Acrobat was then a suite: Acrobat Exchange, for PDF viewing, editing and printing; Acrobat Distiller, for PostScript to PDF conversions, and Acrobat Reader, a utility for PDF viewing and printing.

A hard childhood

Despite initial excitement from a pretty large fraction of industry, from Adobe and from Warnock himself, Acrobat begginings wasn’t as auspicious as expected. Right before launch, other parts of industry couldn’t see its usefulness, calling Acrobat “a solution looking for a problem”. Warnock had trouble even inside Adobe. It’s known that one evening, former vice president of marketing Linda Clarke was complaining and cursing Acrobat out loud because she found hard to use the program (all employees had to use that beta, so it could actually have been hard) when she was trying to get her job done and go home, a task usually easy to do on paper. Warnock got into Clarke’s office and realized that Acrobat was the culprit. He sat down in front of her, and patiently explained his vision and the goal of the project… along the rest of the night [4].

Adobe’s “mistake” for asking a steep price for the Acrobat product family might have been another reason. At launch, Exchange’s price was U$S 195 per user, Distiller had a tag price of U$S 695 for personal use (two computers at most) and U$S 2,495 for a multiuser version [5]. Even for Reader you had to pay U$S 50, just for the ability of opening and print a PDF [6]. This decision didn’t help a lot to a widespread adoption. Internet download times for typical documents were huge, a limitation for a vision centered on document exchange. Besides, other solutions were already available at that time, with their own benefits: Common Ground, from No Hands Software, was cheaper and offered the ability to attach a viewer program inside documents, allowing recipients to open them with no preinstalled software whatsoever.

Time after launch, Acrobat was looking for its own place under the shade of his “big brothers” Photoshop and Illustrator. Nevertheless, despite the fact that Warnock was not thinking in graphic industry when he conceived PDF (as it happened when he designed PostScript), newspapers and ad agencies began to gain interest in PDF as a self-contained, closed format, which guaranteed that images and text remain intact, lowering process error rates and process times of related tasks. Associated Press, through their AdSEND ad delivery system [7], began asking PDF and Acrobat to have more features, when they noticed a steady increase in files delivered as PDF from their clients; even if it was enough for newpapers, it was insufficient for magazines, catalogues and high quality color printing.

On 1994 Acrobat 2.0 was launched, along with version 1.1 of PDF, with support for links to external PDF (previously it was limited to links to pages inside the same document) and support for device independent color, though yet limited to RGB. Adobe started delivering all developer documentation in PDF format, but they also had a big partner on their side: US government adopted PDF as the distribution format for tax forms, vital in every american’s life. Acrobat first revision 2.1 added support for embedding multimedia content, i.e., sound and video.

Graphic world jumps in

Internally known as Amber, Adobe launched Acrobat 3.0 along with PDF 1.2 in November 1996, correcting 1993 “marketing mistake” by offering Reader at no cost, and it became the first version seriously targeted at prepress by adding the following features:

  • Support for OPI 1.3, a workflow based on low resolution images with smaller file sizes, which are replaced with high resolution ones at printing time; this tool had peak usage years ago, when networks were slower and hard disks smaller, but it’s no more used today;
  • CMYK support (finally!);
  • Support for spot colors;
  • Ability to embed process information such as screening functions and overprint settings;
  • Support for plug-ins, allowing third-party developers to add extra functionality to Acrobat.
Acrobat Exchange 3.0, la primera versión "amigable" con la gráfica.
Acrobat Exchange 3.0 splash screen, the first “prepress-friendly” version.

This additions drove PDF adoption across printing industry. For instance, digital black-and-white printing market started to regularly use PDF to print to Xerox digital presses of that time. Between 1997 and 1998 plug-ins architecture allowed companies like Enfocus to build valuable tools for digital prepress: PitStop and CheckUp showed up for the first time as Acrobat plug-ins.

On 1998 Agfa launched its Apogee prepress system, completely PDF-based and targeted at color commercial printing, becoming the first company adopting it in an integral commercial solution. The rest will follow soon afterwards.

On the web side things also started to get better. A new plug-in launched for Netscape, the “browser of the hour”, allowing users to view a PDF inside a web page, leveraged a rise on the format popularity in the early internet. A web page could have a link to a PDF and the other way round, so PDF became a “first-class citizen” on the web.

Acrobat 3 was the first version including Japanese among its available user interface languages, improving its adoption on the other side of the Pacific.

Quest for the rules of the game

At certain point in its history, PostScript “betrayed” itself: its great power as a page description language made it more error-prone. PDF as a “light” version of PostScript doesn’t have that power and therefore cannot suffer from those kind of errors; this fact finally helped to its adoption, replacing PostScript.

Nevertheless, PDF went thru something similar. Once it supported the main features demanded by industry, all other “non printing” additions (such as JavaScript code, multimedia content, to name a few) started to be viewed as potentially dangerous in a document exchange workflow. In other words, format versatility means that only few PDFs (out of possible ones) are suitable for a printing application. What to do then? Go back to PostScript?

Convinced that PDF is per se a good solution, in 1998 several companies started together to work in defining rules that should be enforced when creating a PDF to qualify as a valid exchange format within a specific graphic application. This required to clearly define:

  • Which are the minimum set of application contexts that embrace most interchange cases?
  • For each context, which PDF elements should be mandatory?
  • Which PDF elements should by no means be present?
  • Which PDF elements might optionally be present and which values are allowed?
  • Which PDF elements are recommended to be present, not being mandatory?

These observations built up to a series of recommendations named PDF/X, where the X stands for eXchange. As we’ll see, they turned into ISO standards few years later.

Meanwhile, as these events were unfolding, in April 1999 Adobe launched Acrobat 4 (internal code name: Stout) along with PDF 1.3, which was considered the first full featured version for prepress, with some confusion to its users: Acrobat Exchange became simply Acrobat (as the name of the full suite). It included some interesting features, like a new color space to improve support for spot colors, smoother gradients and bigger page sizes (more that 3′ per side). It was also considered the most buggy version ever: a lot of upset users force Adobe to promise a new 4.05 version with those bugs corrected, but things seemed to go worse when Adobe pretended to charge users for that fix. Finally Adobe listened to them and sent free copies to its registered users, although European clients had to wait up to four months to get it.

At this point it was clear industry had definitely adopted PDF as de facto file format. The road was open.

A farewell to PostScript

Acrobat 5, codenamed “Brazil”, made its debut in May 2001, introducing PDF 1.4 and a break point to designers and printers alike: announced by Adobe a year before through Illustrator 9, this version came with transparency support [8]. From now on it is possible to design objects that keep their opacity setting natively; before this version, as in the PostScript case, any group of objects having transparency should be resolved (i.e. replaced) by opaque elements resulting in the same appearance, a procedure called transparency flattening. At this point Adobe sent a clear signal: PDF is the way to go, having received a feature denied to PostScript.

Acrobat 5 also included the new color engine Photoshop users already enjoyed: Adobe Color Engine or ACE, key piece of the entire color management of Adobe applications so far.

Besides, this is the first Acrobat version with the ability to correctly simulate overprinting on screen (although this feature was disabled by default).

The road to standardization

isoThat 1998 initiative managed to make its way to ISO: in 2001, a group of standards [9] under the denomination ISO 15930 (Graphic technology – Prepress digital data exchange – Use of PDF) appeared for the first time to define what a “correct” PDF should be, and to that end three application contexts were proposed:

  • PDF/X-1, also ISO 15930-4 (Full blind interchange with CMYK data), aka “printer’s paradise” because it demands the file to have all fonts and images embedded, and every element defined in CMYK color space and/or spot colors;
  • PDF/X-3, also ISO 15930-6 (Full blind interchange with colorimetric-defined data), extending PDF/X-1 by allowing elements to be defined in Lab or other color spaces as long as they had a color profile embedded;
  • PDF/X-2, former ISO 15930-5, no longer available (Partial interchange with colorimetric-defined data), extending PDF/X-3 by allowing certain elements not to be present in the document, as long as a workflow to univocally find those elements at the final stages is defined.

All three standards have in common some mandatory features, chiefly definitions of several “boxes” containing the graphic elements (media box, trim box, bleed box), intended printing conditions (by means of a reference by name to a specific standard or by embedding of a color profile), and a flag to indicate if document has already had trapping.

In April 2003 PDF 1.5 and Acrobat 6 were launched, the latter known as “Newport” on development stage. In this version Adobe logically took the next step: besides transparency, Acrobat added support for layers, a feature packaging applications are currently using. Moreover, making use of a successful experience from Acrobat plug-ins developers, Acrobat Professional was launched, including right out of the box valuable prepress tools such as:

  • PDF optimizations: file size reduction, downsampling, deletion of invisible elements and other clean-up actions;
  • Preflight tools;
  • Measure tools (rule, magnifying glass) on the document itself;
  • Transparency flattener;
  • Job ticket creation, needed for the emerging JDF technology, driven by CIP4;
  • Creation and verification of conforming PDF/Xs;
  • Color separations simulation.

As it happened with Acrobat 4, this version was criticized for being slow. Typical of the evolution of many applications, steadily adding new features end up turning a “lightweight”, fast execution program into a heavy, slower one.

Meanwhile, Microsoft requested Global Graphics to develop a new XML-based format intended to “compete” with PDF, named XML Paper Specification (XPS for short) [10]. That move triggered some alarms at Adobe, because XML was already being preferred as a “lingua franca” among internet applications and early Web 2.0, and there was a huge set of tools available to process this kind of format, most of them open source.

PDF, international standard

At this point, improvements to format and software started to be modest. Code-named “Vegas”, Acrobat 7 and PDF 1.6 hit the market in January, 2005. Among a few new features, this version allowed direct embedding of OpenType fonts; before it fonts must be embedded as PostScript Type 1 or TrueType, the accepted font formats. Another feature is the ability to embed a collection of other files (PDF or otherwise). Finally, it allowed embedding of 3D models from a CAD application.

PDF
Standard icon of a PDF document.

Among less friendly aspects (from user point of view at least), this was the first versiòn to require an “activation” process to use, a growing tendency among developers to fight software piracy.

In October 2006 Acrobat 8 was introduced (codenamed “Atlas”) along with the last PDF version controlled by Adobe; about a year later, January 2008, PDF 1.7 became officially ISO 32000-1:2008 [11], although in practice full specifications were published in July 1st.

From this point on, PDF is no longer a proprietary format, last argument from a few remaining detractors (even though its specification was public right from the beginning), but a true international standard, as its development is driven by the industry it serves, and not by a private company.

Adobe might have considered the pros of allowing its format to become an ISO standard, despite the loss of direct control on its development, when they took in consideration Windows Vista launch on fall 2006 and its XPS document creation support, and also Global Graphics launch of the first RIP supporting both PDF and XPS [12].

Adobe attempts a journey to “Mars”

Avoiding to put all eggs in the same basket, in 2006 Adobe started to think about a future, XML-based replacement of PDF: the Mars project [13]. Although it was denied by Adobe, market saw this move as a response to Microsoft XPS format and a fight for the electronic document territory. Nevertheless, general public barely noticed this. There was just a plug-in available in the following, codenamed “Nova” Acrobat 9, launched in June 2008, that allowed to save a PDF in the new Mars format.

Fact is, Mars project evaporated just like ancient water in the martian soil, leaving just another version of Acrobat without a new version of PDF, now controlled by ISO. However, in order to be able to introduce new features, Adobe added to PDF 1.7 a series of “extensions” hoping to merge them later into the standard. One of those extensions allowed for inclusion of geolocation data.

At this point, new layer and transparency capabilities encouraged ISO to extend standard 15930 with two new PDF/X:

  • PDF/X-4, also ISO 15930-7 (blind exchange), may be thought as a PDF/X-3 that allows for inclusion of layers and transparency;
  • PDF/X-5, also ISO 15930-8 (partial exchange), may be thought as a PDF/X-4 that shares the “open” behavior of now “defunct” PDF/X-2.

Acrobat included these new types of PDF/X and offered PDF/X-1a, PDF/X-3 and PDF/X-4 creation and conforming verification. It also included OCR capability for text extracting from scanned documents. At the same time, Adobe PDF Print Engine (APPE) was launched [14], a standard engine for PDF processing, targeted to workflow providers who in the past adopted its PostScript equivalent, the Common PostScript Interpreter (CPSI). At this point, PostScript development formally ended.

Less versions, more standards

PDF-based ISO standards development started to move faster. In September 2010 PDF/VT standard ISO 16612-2 was published, based on PDF/X-4 and PDF/X-5, and targeted to general variable printing workflows; VT stands for Variable and Transactional, with their usual meaning in database technology.

Not just graphic industry benefited from PDF. There are several standards or application guides already published to standardize PDF creation in other contexts [15]:

  • PDF/A (Archiving), defined in ISO 19005, targeted at those documents intended to be preserved in years to come;
  • PDF/E (Engineering), introduced by ISO 24517, describes conditions that documents should met in construction, manufacturing and geospatial industries;
  • PDF/UA (Universal Access), contained in ISO 14289, grants accessibility to disabled people requiring assistive technology;
  • PDF/H (Healthcare), not a standard (yet), conceived for medical record exchange. At present it just consists of a series of good practices guides.

From this point on, maybe it’s not unfair to say that Adobe focused more in business than in shining with a new version. By the end of 2010 Acrobat X made its debut, with a redesigned user interface, followed in October 2012 by Acrobat XI, which ignited a controversy about its new cloud-based licensing model, where users rent the software instead of buying a perpetual license; least, in April 2015 Acrobat DC (Document Cloud) was launched and completed its transition to the new scheme (still offered in both licensing models, contrary to other Adobe applications).

The future of PDF

pdf2.0Specifications for PDF 2.0 are being written right now, which are expected to be published as ISO 32000-2, the first trully post-Adobe standard. Considering that most PDF are still based in PDF 1.4, a valid question arises: Do we need another PDF version?

These are the reasons presented by the industry:

  • Current standards, even approved by ISO, were written by Adobe. Specifications need a neutral writing, where every point of view has been considered;
  • Conformance to clearer specifications reduce software tools developing costs and improves interoperation among applications;
  • Current version 1.7 is based on the Adobe’s concept of “conforming readers” and “conforming writers”. Industry asks for a standard more focused on format itself, instead of applications.

What new features can we expect in PDF 2.0? Maybe not quite a lot, as its primary purpose is to resolve a formal issue, even as a first step to achieve a clearer, worldwide standard. Nevertheless, we already know some intended new features:

  • First of all, more clarity;
  • Improved printing and displaying in all sorts of devices;
  • Printing conditions at page level. This way it will be possible for a PDF to contain a complete job, even if its parts should be printed on different conditions (i.e. cover and inner pages of magazines);
  • Support for color-order specification;
  • Support for spectral colorimetry based on CxF data exchange standard;
  • Full geospatial data support (already introduced by Adobe in previous extensions);
  • Digital signature improvement;
  • Text pronunciation assistance;
  • Support for related files, as introduced by PDF/A-3.

It’s yet unclear whether a new series of PDF/X standards based on this new format will be published. Chances are that PDF/X-4 would be extended to support PDF 2.0 files, but it may be possible that a new PDF/X-6 standard be on its way.

When this new standard is published, it will be interesting to look forward to response from developers and information industry. Anyway, knowing the huge number of companies and organizations of all kinds, private or public, that have done the format their own standard, we can be sure that PDF technology will be among us for years to come.

References

Published inGeneralHistory

Be First to Comment

Leave a Reply

Your email address will not be published. Required fields are marked *