Interview with Microsoft's
Robert O'Hara

The Development Manager for the Windows CE PIMs, e-mail applications, and Entertainment Pack talks about the birth and growth of Windows CE.

Questions asked by Rich Hall, Executive Editor of Handheld PC Magazine

Q: How were you involved with the development of Windows CE?

A: I was involved near the very beginning of the predecessor project to Windows CE, "WinPad." WinPad started in the summer of 1992, I joined Microsoft to work on it in December of that year. WinPad was going to be a machine more like a Newton (portrait mode, handwriting recognition, and no keyboard). It was killed in the fall of 1994, and from its ashes the Windows CE effort began.

I managed the group that wrote the PIM (Calendar, Contacts, and Tasks) and several other applications and components. Personally, I coded the Calculator, Help, and ported the Solitaire game from Windows 95 to Windows CE.

I also created the Microsoft Entertainment Pack for Windows CE, and led the team that produced it. This program, featuring ten games for your Handheld PC, is Microsoft's first retail product for Windows CE.

Moving forward, I also have responsibility for the messaging (e-mail) and wireless applications for Windows CE.

Q: What was it like working on the Windows CE development team?

A: I've been working in the software industry since 1974. (Which makes me the "old man" of this team.) I have to say that this is one of the very best groups I have ever been involved with. The people here are very motivated, and they work very hard. In a word, working on Windows CE has been "intense." Many weekends and evenings were spent here by many people to get the product out. I think it is amazing that we shipped the product in less than two years, when we started from scratch.

An interesting anecdote is how Solitaire came to be in the product. I started porting it in late July of 1995, and had it essentially complete within a couple of weeks. (It would take far less time today, but at that time we did not have the mature development tools of the Software Developers' Kit.) At a project review in August, I got it into the build, and just told people to try it out. Everyone loved it, and from that point forward it was part of the "official" product. By the way, if you want to see who helped make Windows CE, in Solitaire, get the Ace of Hearts on the top row, then hold down (Alt) and tap the Ace.

When we celebrated shipping the ROM image to the manufacturers, I rode my motorcycle through the hallways to let people know we were celebrating. Motorcycles seem much louder in an office than they do outside, by the way.

Robert O'Hara (right, on motorcycle) and members of Windows CE development team celebrate the completion of Windows CE.

Q: When was Windows CE "born?"

A: I mentioned WinPad above. Another project, "Pulsar," was underway at the same time. Pulsar was "a pager on steroids." It was to be a handheld device, with information content delivered wirelessly. But repeated focus groups convinced Microsoft that no one wanted it.

 In the fall of 1994, after both WinPad and Pulsar were reviewed by senior Microsoft management, it was decided to combine the two groups into a single effort which would work to build a mobile computing product. The new project, christened "Pegasus," inherited the WinPad vision: build a mobile device that would be a useful companion to a Windows desktop computer. None of the WinPad software code was carried forward, but the WinPad design continued to influence Pegasus. The Pulsar team had developed a lightweight Win32 subset operating system kernel that ran on a RISC processor. This work continued uninterrupted during the reorganization, and grew into the Windows CE operating system. Through the winter and early spring of 1995 the Windows CE team sorted itself out and developed a product design.

Q: How much did existing handheld platforms influence its design and development?

A: They were a strong influence. We looked at the three most successful handhelds: the HP 200LX, the Psion 3a, and the Sharp Zaurus. They were the target machines, in that they were reasonably successful in the market, and therefore represented something that people found useful and were willing to pay for. If you look at any of the Handheld PCs you can see the strong resemblance to these machines: Clamshell design with keyboard - no handwriting recognition, size, weight, no backlighting (at that time), instant on, battery life in weeks, support for communications, connectivity to the desktop, third-party applications, priced around $500.

We also studied the machines that failed (or were not rousing successes at that time): PenPoint, Momenta, Magic Cap, Newton, and tried to learn from them. Overall, we took a customer- and market-centered approach. We set about to create a pragmatic product that leveraged the largest body of third-party developers and installed base of PCs: the Windows community. We didn't set out to create a "revolutionary" product -- just one that someone using Windows today would find useful.

We took those good ideas, and added some of our own: no-brainer sync with the desktop (this was one of WinPad's key ideas) and strong affinity to Windows, both from the User Interface and the programmer's APIs.

Q: The Windows CE handhelds are positioned as "PC Companions." What was the thinking behind this decision?

A: While many people will use a Handheld PC as their only computer, we believe far more will use it to carry their critical data with them. And that data tends to live on the corporate LAN, accessed through their desktop or laptop PC running Windows. This is part of Bill Gates' "Information At Your Fingertips" vision, something he began talking about in 1990. We focused on the Windows companionship because we understand Windows and what people find useful there. We wanted to work from areas we knew (Windows) to areas in which we were newcomers (handheld computers).

I believe that as Handheld PCs grow in capabilities, with better screens and more memory, more and more people will find they don't need the desktop machine. Certainly we are working on improving the connectivity of Windows CE to make that as easy as possible.

Q: Office 97 came out recently and the press is talking about Windows 97. Connectivity and synchronization and a similar user interface are Windows CE's most important selling points. Originally Windows CE worked with Schedule Plus rather than Office 97's Outlook. How will Microsoft be able to keep the Windows CE operating system up to date without a long lag time?

A: We have an update to the H/PC Explorer available on our Web site. This version fully supports Outlook. For the longer term, we are now working much more closely with the Outlook team, with the Exchange team, and with the Windows 95 team to ensure we track developments as closely as possible. There will always be a little lag, as our product cycles are not in sync. For most people, I don't think this will be a big problem.

Q: What do you like best about Windows CE? What is its most useful feature?

A: There is not really one feature. Rather, it is how easily I can keep information with me that I need while I'm away from my desk. I often download the e-mail I don't get to during the day and read it at home in the evening, responding as appropriate. Then when I connect in the morning I send out all the mail I've answered. I do the same with reports or specifications I need to review: I download them to a PC memory card, then read them in Pocket Word on my Handheld PC. I much prefer to do this in a comfortable chair in my living room than at my desk in the study.

I really enjoy the games of the Entertainment Pack. I've started playing chess while on airplanes (or waiting for them at the gate). It is much more comfortable to sit back and hold a Handheld PC than to work at a laptop in a cramped coach seat. I recently borrowed a book on chess from the library so that I can get better at it.

The other thing is that I've started using wireless e-mail. When you are traveling this is really great. At Microsoft we live on e-mail, so being able to get it wirelessly is like getting a cell phone for the first time. I think wireless e-mail and Handheld PCs are going to be a big hit this year.

Of course, these are things that readers of your sister publication (The HP Palmtop Paper) have been doing for years with devices like the HP 200 LX. The difference for me is that it has never been this easy to move my data between my Windows-based desktop and Handheld PCs.

Q: Microsoft has a wide variety of desktop software. Some of this has been shrunk down/adapted to the H/PC platform and some has not. One thing that seems to be lacking is a database. Why was this seemingly important feature left out and will it be in later versions?

A: Yes, there are several applications "missing" from the first version of the Handheld PC. Along with Access, Power Point has been frequently requested. Why were they left out? It's the usual answer, which is also the truth: we did not have enough time and people to get them done for the first version. This is, of course, true for every project. You never get all the features you want into whatever version you are working on. And of course, in hindsight, the missing feature can be considered by many to be "critical."

Q: What is its most useful, but likely to be overlooked, feature?

A: If you have Microsoft Exchange on your desktop, it is the Inbox Transfer feature. This lets you download mail from Exchange to your Handheld PC, work on it there (read it, reply to messages, forward them, compose new messages), then upload and send everything you did. You don't need a modem, you don't need a remote e-mail account. Now, before you go to a boring meeting you can fill your Handheld PC with e-mail, and read that instead of doodling.

Q: What was the most difficult challenge you had to overcome in developing Windows CE?

A: There were two key challenges. The first was within the group, and that was reaching agreement on a design for the user interface. When we decided to go with the Windows 95 user interface (UI) in the early spring of 1995, it was a watershed decision. Before then there had been major disagreement within the team as to how the UI should look; after that we were just working out the details on how to best adapt the Windows 95 UI to the small screen of a Handheld PC. It was not an easy decision to make, but in hindsight, it was a decision crucial to our success.

The second challenge came from outside of the group, and it is one that we continue to face. This challenge is to overcome the skepticism of the very idea of a handheld computer. Something under one million handheld computers were sold worldwide last year. We, and by this I mean everyone in the handheld computer industry, won't be regarded as a real success unless we can grow that number to several million per year. You can be sure that we on the Windows CE team will work very hard to grow that number.

Of course, the current challenge is to get people to stop playing games during meetings.

Q: Since the Handheld PC is positioned as a Windows 95 companion, why is the Software Developers Kit (SDK) available only to Windows NT 4.0 users, and will future SDK's allow use on a Windows 95 machine?

A: Windows CE is a Unicode system. That is, it uses two bytes to represent each character in the alphabet. This greatly facilitates the creation of versions for the non-Roman alphabets, such as Japanese, Korean, and Chinese. Windows NT is also Unicode, but Windows 95 is not. That is why the SDK only works on NT. Actually, it does work on Windows 95, but you can't use the database calls there. We are exploring ways to make the SDK work on Windows 95.

Q: With all of the feature-rich Microsoft applications available to choose from, how did the team decide on the ones to include on the H/PC?

A: Again, we looked at the existing handheld computers as a starting point. Thus we included a PIM (Calendar, Contacts, Tasks), a clock and a calculator. Then a word processor (Pocket Word) and spreadsheet (Pocket Excel). Because the Handheld PCs do not have color screens, we decided not to include Power Point. We wanted to include a simple database or list manager, but were unable to come up with a design that worked well with the desktop databases, so we dropped that. Initially we did not include an e-mail client, but further market research during the latter half of 1995 convinced us this was needed, so we added the Inbox. We also ranked a list of "future features" in the research that we conducted during the summer of 1995. Automap Streets came out on top. We were able to help the Streets team justify the resources that were needed to create Pocket Automap Streets.

Q: How long was the development process?

A: Twenty-one months. We started in December of 1994, and shipped the ROM images to the manufacturers in September of 1996.

Q: The Handheld PCs are designed to be ROM upgradeable. What new things can we expect to be added to Windows CE?

A: The truth is we don't yet know what new applications will be in the next version. We've received a lot of feedback from Handheld PC users, and we are working to add those features to the product. I suspect that our list of candidate changes would match most users' lists. As the Windows desktop user interface evolves and expands, we will continue to adapt it to the smaller form factor of the Handheld PC.

Q: What's the future of Windows CE (what platforms will it be on, how will it change and improve)? Windows CE is still a new baby. What is this child going to look like five years from now?

A: Some of what will happen is already pretty obvious; various press announcements have been made. Windows CE is being ported to a variety of processors; it has been licensed to many OEMs. Thus in the coming years you will see it (or sometimes not see it) in a variety of products that feature embedded computers. For example, the Windows CE-based Navitel TouchPhone is a desktop phone with a 640x480 screen that provides e-mail and Internet access.

I'll speak personally here, although I believe what I say is shared by much of the team. My dream is that handheld devices running Windows CE will become as common as cellular telephones are today. Bill Gates has spoken of the "Wallet PC" in several speeches. Well, the handheld PC of today is the first small step towards the Wallet PC. I think that the continued shrinking of electronics, the growth of worldwide communications and commerce over the Internet, and the emerging wireless and e-cash infrastructures will make a small, communicating handheld device commonplace within five years.

My fear is to be on an airplane then, when the nice grandma next to me holds up her handheld computer and says, "Would you like to see pictures of my grandchildren? I have over 1,000 of them here!"

Q: Could you tell us a little about the Windows CE book you've written?

A: The book is titled "Introducing Microsoft Windows CE for the Handheld PC," and will be published by Microsoft Press. It will be in most bookstores this May. You'll find it along with the other computer books. The book is aimed at the general reader, not at programmers. It introduces Windows CE and Handheld PCs, then gives a detailed tour of all of the applications. There are a lot of tips and tricks that both novices and power users will find useful. If you haven't purchased a Handheld PC, you might buy the book first to assist you in evaluating it. The neatest information in the book? All of the "Easter Eggs" in the product are described, including how to cheat at Solitaire. Here is a short description:

"Introducing Microsoft Windows CE for the Handheld PC"

This book is aimed at the general reader, not at programmers. It introduces Windows CE and Handheld PCs, then gives a detailed tour of all of the applications. The book includes a lot of tips and tricks that both novices and power users. The book also describes all of the "Easter Eggs" in the product (hidden features and other surprises not described in the documentation).

About the Author: