one of the most interesting talks I've heard in a while was about the effort to rewrite Perl in C++ (i forget the speaker). the most interesting bit of the talk for me was not the choice of language, but the reasoning behind the choice.
one of the main points was that the Perl code is becoming so baroque that the cost of entry is now too high for most people to consider hacking the internals - i.e., it takes too long to understand enough about what's going on to make a contribution.
one of the benefits, i would hope, of a modern OO design would be lowering the cost of entry into the kernel. i note that the design specifically allows people to hack kernel components while the machine is running without affecting other users. this is a major advance and may involve many more people in kernel hacking.
while linus is undoubtably a great coder and has done an amazing job at holding the linux kernel together all these years, one has to wonder how long he can continue to tread the line between keeping a steady rate of progress without the kernel code slowly creeping towards inpenetrability.
Slashdot Mirror
one of the most interesting talks I've heard in a while was about the effort to rewrite Perl in C++ (i forget the speaker). the most interesting bit of the talk for me was not the choice of language, but the reasoning behind the choice.
one of the main points was that the Perl code is becoming so baroque that the cost of entry is now too high for most people to consider hacking the internals - i.e., it takes too long to understand enough about what's going on to make a contribution.
one of the benefits, i would hope, of a modern OO design would be lowering the cost of entry into the kernel. i note that the design specifically allows people to hack kernel components while the machine is running without affecting other users. this is a major advance and may involve many more people in kernel hacking.
while linus is undoubtably a great coder and has done an amazing job at holding the linux kernel together all these years, one has to wonder how long he can continue to tread the line between keeping a steady rate of progress without the kernel code slowly creeping towards inpenetrability.