A senior researcher from Google gave a talk at the compter science colloquium at my institution in December 2005. It was supposed to be about such things as the Google Maps API, but it really was a talk about how great it was to work at Google.
At the time, while doing research for my Strategic Defenestration Initiative (another story), I had read an editorial (URL: http://linux.sys-con.com/read/84967.htm -- the editor could use some editing) in Linux SYS CON which argued that LAMP and open source technology enabled Google and Amazon.com to grow "...quickly into industry leaders." The author pointed out that Amazon and Google were successful partly because LAMP technology allowed them to compete on an equal footing with Fortune 500 companies. And now they were successful enough to replace their open source Linux and Apache infrastructure with proprietary software systems (Windows) if they wanted, but they had not done so.
During the question and answer period, I asked the speaker whether Google had the financial and technical means to migrate their servers from Linux to Windows. He replied that they could do this. I couldn't resist: I then asked the speaker whether Google would go ahead and migrate all their servers to Windows.
His reply brought down the house: "if we did that, 90 percent of our engineers would quit on the spot."
XP is stable. It loads executables, runs them, moves files around etc, without much fuss. Google's code runs mostly on interpreters, so...wtf? The only people who should care are the accountants who would have to pay Microsoft. I mean: Is the punchline, "Ha, ha, google's engineers really are as immature as the bean bag chairs suggest"?
(I withdraw my question if the answer is, "Because they wouldn't want to spend months on useless administrative tasks like re-installing OSes.")
Disclaimer: I haven't programmed on Windows in a very long time, and the only time I did it was in a cocoon of an intro programming class in VS.NET where we didn't have to program anything that was either useful or interesting.
However, one of the more substantive complaints about Windows I've heard (beyond the usual instability and insecurity) is that it isn't easily programmable. You can't (I've read) effortlessly connect small, well-designed, single-purpose programs together with things like pipes, redirection, etc. And, most programs by default (again, I've read) are inextricably attached to their GUIs, so you can't make them scriptable like you can with a command-line Unix tool plus a shell script. That would mightily suck for anyone who's used to the high level of programmability in Unix.
A senior researcher from Google gave a talk at the compter science colloquium at my institution in December 2005. It was supposed to be about such things as the Google Maps API, but it really was a talk about how great it was to work at Google.
At the time, while doing research for my Strategic Defenestration Initiative (another story), I had read an editorial (URL: http://linux.sys-con.com/read/84967.htm -- the editor could use some editing) in Linux SYS CON which argued that LAMP and open source technology enabled Google and Amazon.com to grow "...quickly into industry leaders." The author pointed out that Amazon and Google were successful partly because LAMP technology allowed them to compete on an equal footing with Fortune 500 companies. And now they were successful enough to replace their open source Linux and Apache infrastructure with proprietary software systems (Windows) if they wanted, but they had not done so.
During the question and answer period, I asked the speaker whether Google had the financial and technical means to migrate their servers from Linux to Windows. He replied that they could do this. I couldn't resist: I then asked the speaker whether Google would go ahead and migrate all their servers to Windows.
His reply brought down the house: "if we did that, 90 percent of our engineers would quit on the spot."