In a recent blog post, I asked you to suggest some IT and enterprise software claims you would like me to look into (send me an e-mail or post a comment to get on the list). One of the first questions I received basically asks
"Why so many operating systems?"
The technical reason is that there are so many underlying information systems/processor architectures and that each works best with operating software optimized for it. So different strains of OS emerged for large transaction processing systems that worked well on batches of jobs; interactive processing as is often found in labs; processing loads (often for different companies) where the work would be cut into time slices; mobile devices; embedded devices in machinery; desktop PCs and so forth.
The idea of even thinking of the operating system separate from the device on which it operates is a fairly recent phenomenon. And with appliances emerging, it is likely the IT industry and IT management and staffers such as you readers will revert to again thinking of the device and its operating software as a single unit.
The technical reason was often compounded by business reasons for each manufacturer of the underlying device to write its own operating software. The fact that that explosion of operating software types in the 1970s and 1980s has been winnowed down to a relatively few operating systems is good news for all of us.
A related question was included by the reader:
"There are lots of Linux distributions, but no one standardized Linux variant as (there is with) Windows."
I would say there are actually just a number of suppliers of Linux (Red Hat vs. Novell vs. Canonical vs. the purer open source communities), but really just one variant. The situation is much better than when there were truly different UNIX variants in the market in the 1990s: Sun/AT&T vs DEC/HP/IBM.
In fact, it is the suppliers in the end that accelerated the winnowing process. At first they were happy to let Microsoft own that part of the stack because the margins were low. But as Microsoft started competing with the suppliers for higher margin software, the suppliers decided to counterattack. They chose Linux as the vehicle so as not to repeat the UNIX wars.
And the real reason for so many operating systems is human nature: a combination of the not-invented-here syndrome and the fact that many individuals decide, "Hey, I can build a better mousetrap."
The reader correctly says that millions could be saved "and there can be bigger progress in IT research, in nano and other technologies." The object oriented (OO) approach to software development that began to pick up steam in the 1980s was designed to reach that objective. Maybe it has succeeded and we just don't know it. Perhaps there would now be as many operating systems as there are IT devices (a printer operating system, for example) instead of around a dozen. The better mousetrap syndrome is still a good thing in the long run, even if inefficient and inelegant.
Thanks for the question. Keep them coming.