manager schedule vs maker schedule

Popular comp-sci essayist and lisp hacker extraordinaire Paul Graham recently posted this article on the difference between a manager's schedule and a maker's schedule. This is really inline with my own views on this issue and really sums up a big problem we have where I work with meetings being scheduled with the makers and the impact that has. We've had tons of discussions around the cost of mental context switching, but even that's an understatement of the problem...

Great work, it's always helpful hearing echoes of these types of thoughts beyond my own everyday sphere. Who knows maybe I can use to add some weight to my arguments.

performance tuning to an insane level

Ok, so I have to admit that I've been one to disregard figures around performance when arguing with co-workers over the merit of managed code vs C/C++. I've even used the argument that statically typed languages like Java and C# offer more hints to the compiler that allow for optimizations not possible in unmanaged code. I still have a fairly pragmatic view of the spectrum of cost to deliver (skill set/maintainability) vs performance gains... but regardless of all that..... wow this article completely humbled and inspired me.

I don't know shit.