Dr. Dobb's Journal December 2001
Yes, there really are people who turn to USA Today and other such pap for their daily dose of technology news. Consequently, they live under the assumption that there are only two operating systems at work in the world — anything Windows and something Linux. Too bad. They don't know what they're missing.
From network-centric distributed operating systems like the University of Illinois at Urbana-Champaign's 2K (see http://choices.cs.uiuc.edu/2k/) to tiny real-time systems like Ville Mikael Turjanmaa's Menuet (see page 32), operating systems continue to be fertile ground for research and development. The fascination with operating systems is, in part, wrapped up with power — if you control the OS, you control the system. Applications can be cool, but they can only do what the OS lets them do.
My introduction to operating systems commenced with TRS-DOS, when I went to work as a tech writer for Radio Shack's R&D group. It was there I heard Radio Shack vice president Jon Shirley bemoan how he hated operating systems and hoped never to have to see one again. This was just after some sort of R&D dog-and-pony show demonstrating a network version of TRS-DOS that didn't work the way it was supposed to. Okay, it didn't work at all, as none of the TRS-80 Model IIs and 16s were able to talk to each other. I don't recall why I was even there, except maybe to memorialize the moment for posterity. In any event, Mr. Shirley wasn't there much longer, as he shortly left Radio Shack to become president and chief operating officer of Microsoft, a career move that always seemed odd considering his sentiments concerning operating systems.
Within a few years, however, Microsoft's Windows, IBM's TopView, and a plethora of graphical operating-system front ends stormed onto the scene and I understood Mr. Shirley's point — he wanted all the command-line complexity associated with the DOS READY prompt (the TRS-DOS version of the more familiar C: prompt) to just go away, to be replaced by something graphical. As it turns out, that something was Windows. Finally, it all made sense — at least until I heard current Microsoft vice president Craig Mundy say at this year's O'Reilly Open Source Conference that you can't build a business around operating systems. Go figure.
Mundy's banalities aside, Microsoft is nonetheless doing some really interesting work in operating systems, at least on the research — if not development — side of things. This is evident in a fascinating paper entitled "Operating System Directions for the Next Millennium," written by William Bolosky, Richard Draves, Robert Fitzgerald, Christopher Fraser, Michael Jones, Todd Knoblock, and Rick Rashid, all of Microsoft Research. As its title suggests, the paper (see http://research.microsoft.com/research/sn/Millennium/mgoals.html) lays out a roadmap of what operating systems of the future — Microsoft and otherwise — might well look like. In some ways, what the authors propose is akin to Mr. Shirley's vision, in that the operating system would effectively disappear "so that individual computers, file systems, and networks become unimportant to most computations in the same way that processor registers, disk sectors, and physical pages are today." The overriding characteristic of such operating systems is that they will be distributed, although also self configuring, self monitoring, self tuning, scalable, and fault tolerant. No, this doesn't sound like Ville Turjanmaa's Menuet, but then these future systems won't fit on single floppy either — nor will they even require today's CD-ROMs. While we won't be seeing anything like what Microsoft researchers envision in product form real soon now, the shift will happen (likely in bits and pieces) and we're in for some excitement ahead.
So here's a question that's been asked before, and most certainly will be asked again: If Microsoft has so many really bright people, why does the company do such dumb things? This month's dumb thing comes from the end-user license agreement (EULA) that some pinstriped pinhead added to Microsoft's FrontPage web-site creation software. To whit: "You may not use the Software in connection with any site that disparages Microsoft, MSN, MSNBC, Expedia or their products or services..." Telling customers what they can and cannot do with a product isn't customer focused. What's next? Will the Excel EULA restrict users from creating business plans that might compete with Microsoft? Will the Word EULA prohibit users from writing letters that say naughties about Microsoft? Will Ford and General Motors outlaw drivers from making left turns?
Microsoft should be ashamed of treating its customers in such an unprincipled manner. Alas, shortsighted scams like the FrontPage EULA are in conflict with the vision that researchers and others at Microsoft are working to bring to reality. Apparently, Microsoft sometimes forgets that its customers are pretty smart themselves — too smart to put up with shell games like the FrontPage EULA.
Jonathan Erickson
editor-in-chief
jerickson@ddj.com