Few products have captured as much attention as Microsoft’s Tablet PC (except perhaps Xbox, which coincidently launched the same year). The company’s history working to deliver the form factor goes back to the earliest days of Windows. BillG’s intense focus on handwriting recognition led to one of the first extensions to Windows, Windows for Pen Computing (1992), and a bitter lawsuit with then perceived leader, GO Corp. Subsequently, handwriting recognition was among the very first groups chartered in the new Microsoft Research organization. The small Windows-derived handheld devices, PocketPC, were pen-centric. Then finally in the early 2000s Microsoft began in earnest (meaning in Windows proper) the technology for pen and handwriting that made it into a special edition of Windows XP for Tablet PC and a new type of computer. This arc took place a decade before the iPad was released (purists do love to point out the Newton from Apple came and went in the 1990s). Famously, Steve Jobs tells the story of the iPhone as one that began with a tablet form factor, but that was abandoned in favor of the phone. With such a long story, one could easily write a book on just the evolution of the tablet and the Wikipedia history page represents the industry search for a paradigm-defining product. In this section, we’ll start with the launch of the modern Tablet PC from Microsoft and detail the much more difficult (in my view) challenges of building software for the form factor in the context of Windows.
Back to 071. Resolving NetDocs v. Office
I should have been prepared for what transpired following the unveiling of our Tablet PC. I was not.
The line to enter Bill Gates’s keynote at the November 2001 COMDEX snaked through the hotel and required metal detectors and body searches following the events of September 11, 2001. Themed Digital Decade 2001-2010, it was set to be a highlight of the week—discussing bringing digital innovations across industries and emphasizing increases in productivity, richness of business infrastructure, and significant changes in home entertainment.
Previously at Fall COMDEX 2000, Microsoft unveiled the Tablet PC, showing a prototype tablet created by Microsoft, highlighting BillG’s personal favorites of pen-based handwriting capabilities and online document reading. At the time printing was still the standard for sharing business information. The prototype was how we hoped to convince OEMs to build similar PCs.
A recurring theme for the Windows business was to use big tradeshows to demonstrate exciting new PCs across the PC ecosystem. JeffR joined BillG on stage to show off multiple innovative PCs running the new Windows XP Tablet OS, many of which were to be available later in 2002 with the tablet-capable version of Windows XP. As with all things Windows, the strength in bringing the product to market was emphasized by a broad array of hardware of different sizes, shapes, and price points. Building off the first party prototype demonstrated the previous year, the keynote brought the full strength of the PC ecosystem to this new form factor with support in new variant of Windows XP.
Alan Kay, a pioneer in visualizing and prototyping the concept of a tablet while at Xerox PARC, conceived of the DynaBook, perhaps the original tablet. The DynaBook was envisioned as a “personal and portable information manipulator" described in his amazing paper, A Personal Computer for Children of All Ages (1972). Of the launch of Microsoft’s Tablet PC in 2001, he told Newsweek’s Steven Levy, "Microsoft's Tablet PC [is] the first Dynabook-like computer good enough to criticize." In PARC-speak, that was a high form of praise, not a back-handed compliment.
The quest for a PC tablet was not new and perhaps dates back in our memories to Captain Kirk on the bridge of the starship Enterprise with his tablet and pen. In the 1980s and 90s the PC industry was buzzing with innovative large screen computers based on tablets and pens from the likes of IBM, GRiD, Momenta, NCR, Compaq, and Asian companies like NEC, Samsung, and Toshiba. Small screen computers arrived (and departed) as well. Palm had a breakthrough with its Pilot, while Apple failed with its Newton. Windows CE devices were blessed with a stylus as well.
What Microsoft envisioned to do differently was to build a new class of computer. The Tablet PC would be a full notebook sized computer, but better. It was a notebook PC in power with the convenience of an actual notebook and pen. It was also fully capable of running the latest Windows and Windows software. This was not only the core of the Tablet PC strategy but the core of everything Microsoft was doing with the “scalable Windows architecture”, or “one Windows, with many implementations”.
Microsoft’s new Tablet PC group, led by Alex Loeb (AlexLoeb), reported to JeffR’s division (rather than Windows) to streamline delivery of a killer productivity solution. As the original manager of the pen computing effort a decade or more earlier, JeffR had a longtime passion for pen input (Some personal trivia is that my first trade show booth duty was showing off C++ support for Windows 3.1 for Pen Computing 1.0 at Spring COMDEX 1992). Jeff was an early fan of PocketPC devices, never missing a chance to use the stylus to jot down notes or show off the latest financials in Pocket Excel.
While the connection to information work was clear, this organization structure was a tacit admission that the Tablet PC might require more end-to-end design and implementation than a typical new Windows PC. To build a complete experience required integrating product design and engineering efforts across Windows, Office, research groups where the ink and handwriting technology was being developed, as well as hardware engineering for digitizers and working with PC makers. Coordinating marketing and partnering with OEMs introduced another aspect of the end-to-end effort.
Technology for digital ink progressed significantly, primarily as a result of the increasing focus on screens and digitizers—the technology that picks up a signal of some form from a pen and converts it to a series of datapoints representing what is drawn on the screen. Flat panel displays for notebooks were progressing quickly. The Tablet PC was timed well to capitalize on these innovations as display and touch/ink sensor-equipped panels proved to be the limiting factor for building even a marginally acceptable experience.
From its earliest days, Microsoft Research was working hard to make handwriting recognition a reality. Handwriting recognition had long been one of the fundamental linguistic technologies (along with language translation and bi-directional speech) that were always 5 years away (since the mid-1950s). Our state-of-the-art approach was fascinating in hindsight. Improving digitizers were able to capture many more points (coordinates) as the pen moved across the screen, even when moved quickly. These points were used to connect dots, which could then be smoothed out to look like a continuous line of ink. Meanwhile, the starting point, ending point, direction, and other data were used to guess the letter being drawn. Recognition was obtained by comparing …