Richard Coleman <coleman@math.gatech.edu> writes:
Although many will not agree with this (and this is just my opinion),
the Guile project is the last/best hope for the Scheme community. It is
the only Scheme implementation that could have a high enough profile
(due to FSF) to gather enough support to add all the necessary extras.
I really hope this happens.
I have given up on the elders in the Scheme community. They have never
had the practicality of Larry Wall or John Ousterhout. Guile has been
going off in the wrong direction, making choices based on GPL politics.
Although many will not agree with this (and this is just my opinion),
Cygnus's Kawa is the last/best hope for the Scheme community. It is the
only Scheme implementation that could have a chance to add all the
necessary extras to gather enough support in the form of a user
community. I really hope this happens.
How can Kawa accomplish this? By leveraging industry commitment to Java.
Java 1.1 Reflection will allow Scheme to access many system services
that have been lacking in a single implementation in the past such as
networking, subprocess manipulation, databases, user interfaces, etc.
Kawa also gains a lot more by compiling to Java byte code. What other
language has so many runtimes competing for the best JIT and garbage
collection? How many runtimes offer portability to Windows, Unix, and
Macintosh, in decreasing order of importance...
Having worked with Olin Shivers on the Scheme Shell and having done my
own embedded Scheme interpreter in Java I know the power of combining an
elegant language with a system programming APIs. I would encourage
everyone to take a look.
http://www.cygnus.com/~bothner/kawa.html
-bri
|