From: | Darius Blasband <darius@phidani.be> |
Newsgroups: | comp.compilers |
Date: | 23 Jan 1996 17:37:13 -0500 |
Organization: | EUnet Belgium, Leuven, Belgium |
References: | 96-01-037 96-01-058 |
Keywords: | OOP, interpreter |
John Gough <gough@dstc.qut.edu.au> wrote:
> There are two additional problems which need to be solved. [Useful
> Java programming depends on its base classes, and the array
> implementation uses dope vectors, different from most compiled
> languages]
> So, in summary: I think that a subset of most languages could use
> java bytecodes as an intermediate form. However, the translation of
> the whole of even suitable languages to java is challenging but not
> impossible IMHO.
I guess that it does depend badly on the kind of language one wishes
to compile. We do have a project here to compile our home-made
language (called YAFL) into Java byte code.
You might say that YAFL is close to Eiffel, restricted to single
inheritance, with extensive and highly typed genericity, and with an
array semantics amazingly close to Java's. Hence, it should not be
extremely difficult to generate Java byte code. The nice thing about
this is that it would even allow us to mix natively compiled YAFL code
with Java byte code dynamically... Funny stuff... However, I guess you
would be right in saying that we are just very lucky people: YAFL's
semantics is pretty simple to map to Java's byte code. I would not
dare consider compiling Fortran or Cobol this way...
Darius
--
Return to the
comp.compilers page.
Search the
comp.compilers archives again.