On Wed, November 21, 2007 07:09, Waldemar Kornewald wrote: > On Nov 21, 2007 2:34 PM, "S.J.Chun" <[EMAIL PROTECTED]> wrote: >> For me, the squeak environment - image - is not problem; but the >> performance >> is my main obstacle to use it for main development. If squeak can show >> more than 90% >> of performance of C in every case(this is important part, for some >> limited area it might >> already performant), for me, squeak is best choice. I hope >> PEPSI/COLA/etc could help >> me to solve this problem - performance, raw performance. Without this, I >> think that >> the language or development environment cannot be universal tool like C. >> (Yes, I know >> human productivity/performance is more valuable but productive, >> performant and fast >> language is better than productive one with moderate speed, isn't it?) > > To me, it depends. If speed is fully acceptable for 90% of my > use-cases then I'd happily combine that productivity gain with a > lower-level language for the other 10% because then I'm still many > times more productive than with just C. Of course, this isn't very > nice, either, and it would be much better if the programming language > had some kind of mode to describe highly efficient code (same syntax, > but a few restrictions) or to let the compiler do a very deep > optimization (whole-program analysis?) for some parts of your code > where speed really matters. Just a thought. I'm not a compiler > developer, so I don't know how much of this is possible. > > Bye, > Waldemar Kornewald >
I could get used to the Squeak environment if sufficiently motivated. I'm immersed in the Unix world view that everything is a file. I'm currently using Python with NumPy and C where needed for performance. A productive high level language that be used from the metal up to exploratory scripting is attractive. The Canon Cat (http://en.wikipedia.org/wiki/Canon_Cat) was an interesting example of what can be done working from the metal up with a productive language. Some might not consider Forth to be productive, but it worked for them. # Steve _______________________________________________ fonc mailing list [email protected] http://vpri.org/mailman/listinfo/fonc
