On Sunday 13 December 2009, David Brownell wrote: > I admit that when I first came across it, the "don't check for NULL" > philosophy seemed pretty wrong. But, quite a few years later on, now > I see that it's quite effective. The "it's wrong" argument is on micro > scales. The "it's right" one is macro scale. For every small micro win, > there are many more large macro wins.
And some similar situations, somewhat counter-intuitive results that are interesting (IMO) but not relevant here: - A professor of mine once collected scenarios where you could come up with highly accurate algorithms ... but if you were willing to compromise a bit and *STILL* fail less often than the hardware used to run the algorithm (!!), you could get significantly better performance. O(N) vs O(N^4), etc. - I recently scanned a document about FPGA usage in space systems. It turns out they needed to make VHDL/Verilog compilers avoid certain optimizations ... like, if you code redundant or voting schemes to cope with alpha particles inducing errors, you really don't want the compiler to take out that redundancy, and increase your failure rates! One of the common threads there is needing to have a handle on just how common an error is before choosing strategies for dealing with it. - Dave _______________________________________________ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
