2011-07-20 ``Language Shootout'' is a term I learned just some days ago. It describes a comparision of programming languages, a benchmark you could say. ``Benchmark'' implies that the comparision is focused on speed and size. Unfortunately exactly that is true for most such comparisions. They focus almost only on the hard facts. But those are pretty much irrelevant and will become even more irrelevant in the future. What really matters are the soft facts: The clarity of a language, the development speed, how error-prone it is, how much code you need to solve a problem, how maintainable it will be afterwards, how much learning time is re- quired, how good it is suited to solve a particulary problem, and so on. These are the things that are seldom measured ... sure they're difficult to measure, but they would provide real value. Execution speed is so irrelevant! Why do people care so much about it, then? The situation is as if people would compare cars only by maximum speed and the width of the wheel track. Michi sent me a link to a nice list of implementations of echo(1). [0] System V5 vs. GNU coreutils and many more. The result is easy to guess, of course. Although most would favor the smaller and cleaner code, they still like to use the better featured version in use. This is the point where I differ: I care about good implementations and thus care to use such pro- grams. Unfortunately somethings the price to pay is too high. ... if users would need to understand the source code of the web browser they use. No, web technology designers should need to do this first. Linux, the kernel, can display a boot logo, which usually is a small Tux, one for each available CPU. Some time ago I applied a patch to exchange the standard Tux with the Crux Tux. Now I de- cided that there is no reason why I shouldn't have a completely custom boot logo. Hence I created one with Gimp and converted it with netpbm: pngtopnm logo.png | ppmquant -fs 223 | pnmtoplainpnm >logo.ppm (Found this command somewhere in the web.) [0] http://gist.github.com/1091803 http://marmaro.de/lue/ markus schnalke