The PC at 25: If I Had a Time Machine, I Would Make One Small Change

On August 12, 1981, I was a pimply faced teenager who knew computers existed but who did not have the cash to acquire the new IBM Personal Computer that was launched that day. I didn’t even have a clue that I would end up in the computer business. Because I was a space buff, I knew about IBM, which designed so many systems for NASA. But the IBM PC–soon to be in my high school, sitting right beside the TRS 80, Commodore, Atari, and Apple in the computer room–is what made IBM a household name. If you want to pick a day when the computer business changed, that day in August is probably the most important one.

That was the day when Big Blue slapped its seal of approval–its own IBM moniker–on the idea that little computers and real people had a place in the computing world. IBM didn’t invent this idea; in fact it resisted it as far as and as long as possible. But once IBM jumped into the PC market, the PC market took off and changed the world. And thanks to a happy accident–an accident mostly because of the ignorance of the mainframe-centric top brass at IBM as to what the “wild ducks” who ran the IBM PC unit in Boca Raton, Florida, were up to–the IBM PC was an extremely open architecture. And that fact, more than any other, accounts for why 1.6 billion IBM-style PCs have shipped in the past 25 years and why most of us sit at an offshoot of that original IBM design.

Intel, which supplied the 8088 chip at the heart of the IBM PC, and Microsoft, which threw together a quick and dirty MS-DOS operating system, often get a lot of credit for making the PC business what it is today, but they were mainly beneficiaries of the initial openness that IBM mistakenly (from the controlling point of view of Big Blue, at least) wove into the first Personal Computer. Once that genie was out of the bottle, it could not be crammed back in, and heaven knows IBM tried to do so with the PS/2 platform, its Micro Channel Architecture, and the OS/2 operating system it created with Microsoft back in 1987. IBM lost that hardware battle to Compaq, lost the software battle to Microsoft (which arguably set IBM up to fail by coding OS/2 in assembler and Windows in C), and having lost those two battles, it lost the PC business. One could argue that, in reality, IBM never had it in the first place.

Because of the 25th anniversary of the PC this week, pundits, analysts, and other people who have an opinion are making comments about what could have been or should have been. Some people think that the Apple designs, which were more faithful to the concepts created at the Xerox PARC labs, should have been the dominant PC platform. Others think that Digital Research’s CPM operating system should have been anointed on the PC, since it was technically more slick than the PC-DOS variant of MS-DOS that Bill Gates slapped together at Microsoft for the IBM PC. I have thought about the origins of the PC a great deal in the past two decades. And I want to toss in my two cents.

If I had a time machine, I would make one small change for that day. As much as I like and respect Gary Kildall–who created CPM, who created some of the core pre-emptive multitasking and windowing technologies for operating systems, and who died in 1994 in relative obscurity–I would not make CPM the operating system that IBM chose, as many would. I want the IBMers who were responsible for picking the MS-DOS operating system to miss their flight and not meet Microsoft, which was working on the Basic compilers for the PC. I want these IBMers to somehow run into Bill Joy, who was at the University of California at Berkeley, putting the finishing touches on the open source BSD 4.1 Unix operating system. (Joy is one of the founders of Sun Microsystems, among many other things.) In October 1980, the Defense Advanced Research Projects Agency (DARPA), the same military organization that funded the basic technologies behind the Internet today, created the Computer Systems Research Group at UC Berkeley. And I can imagine a world where BSD Unix, which was technically more elegant than a lot of other operating systems back in 1981 and is still better than most today, is on the PC. It is not hard to imagine. BSD is at the heart of Apple’s Mac OS X operating system, in fact, and has been since 2001.

This seemingly minor switch from MS-DOS to BSD would totally change the facts of computing and the forces in the computer business. First of all, the Internet might have become commercialized sooner, since BSD was the early ARPAnet. And with IBM behind BSD, AT&T might not have launched into a lawsuit against UC Berkeley that significantly slowed down the development of open source software, BSD Unix, and the deployment of Internet technologies. BSD probably has the most generous and useful open source licensing terms in the IT industry. It is open without being as invasive and viral as the GNU General Public License that was created by open source guru, Richard Stallman, of MIT. So maybe we get the Internet and the open source software revolution just after the launch of the IBM PC, itself an open platform by accident more than design, but open nonetheless. Intel still gets stinking rich, but Bill Gates doesn’t unless he creates a very good, commercialized implementation of BSD. Because BSD is open source, even if Gates creates a Windows operatign system based on BSD, he can’t charge a lot for it, since he can only charge for the value he adds and anyone else can add similar value and charge a similar price. Open source means there can be no desktop Windows monopoly to build and maintain in the first place.

With BSD being the clear leader on the desktop, it moves into servers, a lot earlier than AIX, Solaris, Tru64 Unix, HP-UX, and a long line of other Unixes do, obviating the need for these incompatible Unixes, and thus we avoid the wars between Unix International and the Open Software Foundation. There is no need to unify Unix because it never splits–at least no more than the three major BSDs have today. MS-DOS doesn’t happen, and neither does NetWare, Windows, or Linux. We all grow up with a secure, stable BSD computer that is a lot harder for someone to hack, and we all grow up with open source software as a given, not as a revolution. Bill Joy is the hero of operating systems, and maybe Bill Gates ends up being a perfectly fine compiler writer for Borland.

And Frank Soltis still does his PhD thesis on the System/38 architecture years before the PC is launched and the System/38 comes to market in 1980. IBM still creates the System/36 and AS/400 lines and the OS/400 platform develops much along the same lines as it did. The AS/400 never was mainstream, so maybe IBM props up OS/400 onto a BSD kernel in 1992 or 1993 and creates a platform very much like Mac OS X, but with an integrated database and application development toolset, like the current i5/OS platform, and it has an integrated graphical user interface and tools that think graphically, too, like Mac OS X. And having done so, the AS/400–since no one is stupid enough to change its name three times–has a 20 percent share of a $50 billion server market. Which means all of us in OS/400 Land are doing better.