Recently I attended Microsoft's annual Systems Software Seminar. This is less a seminar in the academic sense than a carefully orchestrated presentation for the press, a gathering where Microsoftians like Bill Gates and Steve Ballmer take turns at a podium and argue over exactly when OS/2 will take over the known universe. There were also presentations on future enhancements to DOS, the OS/2 LAN and SQL servers, future developer goodies, etc.
Bill Gates remarks on programming languages really got me thinking, though. He spent a fair bit of time explaining to an audience of industry analysts and press people about these things called "objects" and how they were the coming thing in programming. This is a view I quite agreed with. The increasing complexity of environments like the Mac Toolbox and the OS/2 Presentation Manager cry out for new progrnmming tools and methods. Language packages like Actor and Smalltalk/V have the potential (in their inevitable protected mode versions) to be the next wave in programming.
But Bill Gates wasn't talking about Smalltalk or Actor.
The people at Microsoft, you see, have a language division that's expected to turn a profit. It isn't their job to convert the world to new languages, it's to sell language packages. Preferably more than Borland does.
So when asked how Microsoft was going to address the topic of object-oriented prograrnming with its languages, Bill Gates wasn't interested in weirdo languages like Smalltalk; he talked about adding object-oriented extensions to existing Microsoft languages: C and BASIC.
Object-oriented BASIC! What a concept! Soon beginners will be able to add modular meatballs to their spaghetti code. My mind boggled with the possibilities and almost missed the follow-up from Gates: there's no promise of when these extensions will appear, or in what form. ANSI C is still the chosen language, and converting to C++ seems unlikely.
The next day on the plane home I was still pondering the notion of what--for want of a better name--I'm calling OOBASIC (pronounced "uh-oh BASIC"). The question I was wrestling with is non-trivial: how much of the increased sales of programming languages is based on advances in programming technology, and how much is simply improved marketing? It's a difficult question. Alter all, Borland had a Modula-2 compiler years ago and dumped it. Yet the same company has managed to make money on PROLOG by hyping it as "the language of artificial intelligence."
I've spent some time considering Bill Gates' opinion that no new languages will make an impact in the years to come, and I'm still not convinced. It seems to me that the new paradigms and new environments will demand new approaches, not patches on old languages. The folks at the Whitewater Group and Digitalk and other companies seem to be able to do quite well marketing solutions rather than trying to jump on the bandwagon and sell yet another ANSI C compiler. Just maybe they're on to something. And maybe object-oriented extensions to older languages are really a step back instead of forward. The next year or so will tell.
In the meantime, I'll be spending my evenings exploring Smalltalk and other languages, examining new paradigms, and counting myself lucky on the days I don't get a press release announcing object-oriented COBOL.
Tyler Sperry
editor