Let me preface this with a small bit of history. I've had apple computers since the screens only had one shade, green. With that said, I actually don't have a mac at home now and I'm typing this on windows 7.
I've always had issues with how Microsoft released their operating systems. If they WANT people to upgrade (and trust me, they WANT people to upgrade), then don't freaking charge multiple hundreds of dollars for your operating system. I mean jesus, the majority of the PCs on the planet use freaking windows. Do you really need to charge $300 for it? (XP used to be close to $300 IIRC) I mean jeeze, even charging $100 for it seems a bit excessive.
Apple, on the other hand, have been offering their OSs for $30, $40, free? Even in college, when 10.6 came out... it was a no brainier for me to buy it for $30. I didn't have a job in college, I literally had no money, yet I still upgraded.
Look, Apple does shitty things too. I "upgraded" to iOS 7 on my iPhone 4. Yeah, I was pissed. I was SO PISSED that I switched to android for a few months. I realized that the android OS sucks more than the "upgraded" iOS, so I switched back. Don't get me wrong, I'm an advanced user that could actually take advantage of the features on android BUT, in the end, I never used it. Every company does shitty things. I just think apple does less shitty things. (Plus every $300 crap box that people love to buy is just loaded with windows (an OS it can barely run), which puts a bad taste in my mouth.)
Anyway, unfortunately windows is a necessity in many jobs today. Many scientific programs only run on windows. Instrument companies don't have the money to develop their programs for mac and pc (and definitely not linux.) Hell, we have computers in my lab that I had to UPGRADE to XP. (Within the last year.) However, that's changing, and I think that is for the better. I'd love to see scientific programs written for linux honestly. (Please don't write them for mac. I don't want a $1000 computer a few feet away from my 800*C apparatus. Apple will (hopefully) never try to enter the "economic" computer market, therefore they'll never be viable for things like lab computers.)