> Actually you're right : the C paradigm is straightforward and perfect to > handle system code.
there you go. as unfortunate as that is, it is still true. Sometimes, I am very surprised C is still around for so long, and essentially unchanged. But its limitations are well known. > What I meant to criticize is its grammar : it sucks that we cannot parse C > with a simple program. It's not KISS at all and we all pay the price when > checking for errors and everytime we wish to process these weird formatted > text files. > > On the other hand, languages based on s-expressions have this unique feature > of being very easy to parse and can be processed as regular data, every > aspect of the syntax being clean and regular. The problem is that these > languages were historically used by pretentious folks on rare supercomputers > and thus were not developed to fit the low level paradigm very well. Allegro just recently got SMT/SMP, all others are probably (no research, just throwing it out there) still stuck in single land. While the folks on the supercomputers today are running SMT/SMP in their zillions. That's the cool thing for them and your s-expressions won't help them today. > This does not have to be. We should bridge this gap that was forced by ice > age investors. Of course i really don't know how nor how it relates to > OpenBSD, apart from this common goal of achieving correctness. Just yesterday, I tried chicken scheme and compiled it to C code. The C was unintelligible, why? Because it is setting up the JVM like environment. This goes back to the vmmap thread. Many people try to play OS designers, when very few of them are capable of it. Just leave the OS stuff to the OS people, they will get it right eventually.
