language interpreters/ interpreted languages weaknesses?

[Michael Hudson]

In an effort to point out just how silly this whole thread is, I'll just
mention that if there was such a thing as a black and white distinction
between interpreted and compiled, it would be a property of the
implementaion, not of the language. Yes, some languages are designed to be
compiled, some to be interpreted. 

But trying to draw a line down the middle of the world and assigning
things to one side or the other is a childish pursuit. 

For starter's: what about lisp?

In a lisp system a function you just type in will usually be interpreted,
but then at a later date can be compiled to native machine code.

On Fri, 3 Sep 1999, phil wrote:

> Yes,
> i don't understand the logic here. A compiled language uses a compiler
> to produce machine instructions and requires no further assistance
> from language specific tools.,

Hmm. Lisp (in some implementations, e.g. cmucl) is most certainly compiled
for any reasonable definition, but does not have a separate compiler, nor
does it "require no further assistance from language specific tools" as
the compiled code still needs the runtime around (for gc if nothing else).

> C, C++, Cobol, Fortran etc. (in most implementations) are the former.
> Python, Java etc. are the latter.

What about gcj (Cnu compiler for java)? Compiles java to native
executables.

> JIT is an attempt to make Java into the former , SI
> (SpanishInquisition) is an attempt to make Python the same. (SI
> doesn't exist but nobody expected it, anyway).

Good line!

Regards,
Michael