[REBOL] curiosity killed the code-generator Re:
From: bhandley:zip:au at: 6-Aug-2000 12:56
I've thought about this myself a little, just for curiosity.
I suppose if you're a Java developer then you should be able to come up with
a few situations, like help with implementing design patterns say. This then
is at the level of manipulating and interpreting Java source code.
At a different level. Another thing you could do is leave the Java source
behind and look at the Java Virtual Machine itself. In some ways this may be
simpler. You could create a Rebol program, no doubt using a particular
dialect, that would generate class files that could be directly executed on
the JVM. If you like a feature of the JVM or some library of functionality
already implemented for the JVM, this could be a way to leverage it.
Obviously these languages have their strong and weak points.
The thing is, I may be a little "uncreative" but I came to the conclusion
that at the moment, unless there is a specific feature, "just" doing it
Rebol is probably better.
Brett.
----- Original Message -----
From: <[balayo--mindspring--com]>
To: <[list--rebol--com]>
Sent: Sunday, August 06, 2000 4:15 AM
Subject: [REBOL] curiosity killed the code-generator
> howdy list,
>
> this one falls into the "I'm just curious" catagory. What would be some
of the benefits of writing a code-generator in REBOL, say, for Java? That
was actually suggested to me some time ago, and a person I mentioned it to
called it a "trial by fire". hmm. Someone wrote some fortran tools, didn't
they?