Saturday, May 17, 2008

Competition v. Monopoly in I.T. -- it doesn't work the way they teach it in history books

One of the most important virtues in western capitalism is supposed to be competition. A lot of legal history deals with this, such as the anti-trust laws that one studies in American history, and that one sees mentioned all the time today when corporate mergers are proposed.

So, I wonder about how big monopolies evolved in the computer business. By the late 1970s, IBM had almost complete domination of the commercial mainframe “general purpose” computer business as far as operating systems were concerned. Univac, Burroughs, Data General, NCR, and RCA had all tried and been merged out of existence. RCA had emulated IBM’s languages but had different “JCL”; Data General eventually became important in the mini market. It’s true, IBM had various operating platforms: VM, DOS, MVS leading eventually to OD-390 and the ability to emulate other platforns. And other manufactures (Amdahl and Hewlitt-Packard) made mainframes that copied the IBM operating environment.

Why did this happen? This was not like the AT&T “regulated monopoly” of the phone business of the 1970s. Part of the reason is that IBM came to control the job market. Employing managers came to expect IBM experience in the various areas (including CICS and IMS), and programmers, to remain marketable, could no longer afford to build their careers around competitors. It is true that in the consulting business there was plenty of competition, but EDS led the pack for years. By the 90s, most major commercial shops had come to rely on mainframe software packages (like Vantage for life insurance, after the older VLN failed to thrive) and the job market started to expect detailed expertise in these packages.

But the market broke apart, starting in the late 80s (ironically in a business environment filled with hostile takeovers), as minis and the PC’s became more important, and then in the 90s with the rapid advance of the Internet. (At the same time, there was a spike in demand for old-school mainframe programmers to handle the phantom Y2K “problem”). But once again, one company started to dominate things: Microsoft. Apple would run a distant second, and some programmers would develop a theory that everyone should deploy Linux on PC’s. (Apple actually offers that in terminal mode but has its own proprietary scripting language.) Even with its monopolistic posture, Microsoft found it in its best interest of offer properly licensed software to PC’s for low prices that most hobbyists with a middle class income could afford. Some people think that the search engine market is now monopolized (by the company that hosts these blogs), but in such an arrangement, companies have offered consumers tremendous content and functional resources (including self-publishing and entrepreneurial income-generating opportunities), often for free, or often for minimal investment.

There are fewer PC manufacturers than there used to be (remember all the way back to TRS80, Atari, Osborn, Commodore, etc?) and only two or three major paradigms, yet prices, even with a falling dollar and international issues, seem to drop relative to capacity. Healthy competition in database software seems to exist (Oracle, Sybase, DB2, etc) partly because a lot of innovation continues and remains to be done. The same is true somewhat in software engineering and languages (java from Sun, C# from Microsoft, etc.) And recently, Microsoft has announced that it would participate in providing laptops to children in developing countries, with more announcements due as to just how they would work technically.

So on the information highway, the concept of “monopoly” doesn’t operate quite the way they teach it in history books.

No comments: