Im my infinite naiveté, I always assumed that developers strive for the better. Usability, performance, stability, whatever: I reasoned that they want to improve it, and the many obvious examples to the contrary just led me to believe that developers, like ordinary humans, can fail.
Fact is, of course, that developers are apparently sometimes guided by entirely different principles. Well, judge for yourself.
In former times, I used to utilize the calculator available on almost every computer as a benchmark tool. It allows straightforward comparisons of the CPU performance of different machines with the same OS, or comparisons of the algorithms employed in different OS on the same machine.
For example, the calculation of 100000! required about 100 s with calc 5.1 on Windows XP SP3, but only 4 s on KDE 3.5.10 with kcalc 2.0.6 on the same system. Obviously, the algorithm used for the latter is the better one.
Now, on KDE 4.2.0, kcalc 2.4 zaps through 100000! in the blink of an eye, 400000! in 0.3 s, and 1000000! in 1.5 s. Its predecessor can process 400000! in 100 s, but I'd rather not try the feat of calculating the faculty of one million there. *g* In other words, the developers of kcalc have managed to speed up the calculations by two orders of magnitude. Not too shabby.
I'd expected a similar development for Windows 7, particularly after the calculator therein was boasted by blogs and official media alike as being amongst the most remarkable features of Windows 7.
1000000!: Overflow, 100000!: Overflow, 10000!: Overflow, 1000!: Hey! Result!
I can't help but being a little disappointed.