In terms of my own aggro threshold, I will always take the thing that works more often, even if the end result is less efficient internally.

I mean, a failure means I spend an unknown amount of time fixing it. There may be a case to be made that the time you save using things that work faster in the first place outweighs the time you spend fixing what's broken, but I will live with a 30% productivity hit on a system that works 95% of the time long before I decide to move to a system that removes the hit but only works 80% of the time.

Sure, the latter works a lot better overall, but I know when I'm sat there fixing a problem and I'm not especially aware of time lost to an inefficient use of hardware.

I'm sure lots of folks would chime in and say if only Windows worked like you want it to 95% of the time. I'm citing random numbers for the sake of argument.

I would love to move to linux. If I knew I could drop an install of it on my machine and know my hardware would work, know my software would work I would do it right now. The fact is, no linux distro can offer that with even 80% certainty. Nobody in their right mind would ever turn around and say everything works with it and your old stuff will just plug in and work.

And by know it'd work, I mean without a fuss. Windows may be quirky, run by The Man and inefficient, but for now I don't care because I know what to expect. I know that I can put windows on a machine and my hardware will be supported, my software will have reasonable tech support, and my peripherals won't suddenly be paper weights.

It's unlikely all that would go wrong in one go by switching to linux, but the uncertainty is enough for me to live with an OS I don't like and don't want to support. My time is valuable enough for me to not want to waste it fixing what isn't broken.