I understand your point and I agree that JVMs should work out of the box, but as it is now, they don't. Your argument is well made and IMHO we should file bugs against existing JVMs that exhibit non-preferred behaviour.
However, in the meantime - as someone who packages a Java *program* - I find myself having to either use these scripts or else tie my app to a specific JVM (which hurts the user). So until the individual JVMs play nicely, and given that debian's priority is its users (not its developers), I'm personally happy to have these scripts around to smooth things over in the meantime and let the user pick a JVM, stick with it and have individual Java apps run properly. > Your script punishes the Debian developer who creates a clean > package for a new JVM, that registers an alternative for > /usr/bin/java and runs without any special help. Users should be > able to install that package and have it just work. Why does the script punish said Debian developer? The first JVM it looks for is /usr/bin/java and default behaviour is not to provide any special help. Thus if the JVM packager does as you say, this script will choose that JVM and work perfectly with it. I don't see your problem. Ben.
