I'm supposed to write 600 words on operating systems. Well, I only have two. Buy Windows. That's it you can go now. Actually I could have written Linux there instead, but frankly I don't care all that much. It's not my job to care about operating systems any more.
I run an IT department, not a computer magazine.
To be sure I used to care a lot about which operating system to choose and was a loud figure in the MS-IBM OS wars, even if I was on the wrong side (that was OS/2. Ed.). I reject the notion that I should choose a bit of software as some sort of anti-capitalist protest against the flagrant misbehaviour of Microsoft. Maybe they are bad people, but they've never done anything to me and I have far more respect for the integrity of the MS people I meet than some of the Oracle reps I've met.
No rational decision I make about system architecture starts with the operating system, indeed I don't really make decisions about them at all. Almost every single time, I simply don't have much in the way of options. We use Exchange, so the choice is which version of Windows Server and no flavour of Unix can be considered.
Where we develop our own code the decision of Linux, HP/UX or Windows is often decisively settled by the presence, or absence, of a key feature. I get worried when I read of people with pure open source strategies or who commit blindly to whatever Microsoft feels like throwing at them. Hands up those who committed to Software Assurance? Keep your hands up if you're not trying to hide the fact and hope no-one notices?
Don't see many hands.
That was a bad decision and frankly I don't have any sympathy for you. When any vendor, be they for desktop PCs, routers or toner cartridges says "buy it now because the price later will be higher", it's a scam. There are no variables in that expression. MS is no worse than other vendors, indeed Microsoft's SA has made me feel happier about myself since I can now feel smug about being smarter than many other people at my level who really should have known better. So I see my job as making decisions, hopefully good ones, but they're my decisions, not those of an accountant at MS, or an evangelist for copyleft.
Sometimes we do have a free choice of what to run a service on. Quite rarely, but it does happen. So I'm supposed to care now? Perhaps, but I'd rather not.
We have some MySQL doing useful work. I asked the man doing the development what he wanted to use, expecting him to say Linux, given the perceived synergy of the open source platforms.
He chose Windows.
His argument was eloquent and compelling. No talk of freedom, security, threading models or multiple platform support. He simply couldn't be bothered to learn Linux. He knew Windows, warts and all, and knew he could bend it to his will.
By now, you may be asking about licence costs and, yes, I care about them. But do the numbers. People cost far more than software. A skilled City techie, spending a month or two mastering Linux, will cost more than the server hardware and software put together. Worse, his first efforts on the new system won't be as good. I'd be paying more for less and, since I work for brokers, such behaviour is frowned on. Nothing is more likely to make a project go badly wrong than forcing the developers to use the wrong system because of some policy formed on a golf course with a tame vendor. Cynically, it also removes a big excuse for non-delivery, since they can't blame you for stopping them.
Some rabid Linux people will howl about security and I care a lot about this. Too much to leave it in the hands of MS or Linux developers. If you're solely relying upon your operating system for security you're already lost. A badly configured server is a risk, regardless of which software you run on it.
Important systems are too complex to be driven by structural prejudice dressed up as information strategy. You have to make informed decisions and if you're always making the same one, the odds are that you're making the wrong one.
Dominic Connor is head of IT for King & Shaxson, a major City broker.