Lisp, Java, and C++

[Scott L. Burson]

   From: Chris Bitmead uid <chris.bitmead@alcatel.com.au> (x22068)
   Date: Tue, 27 May 1997 15:03:13 +1000

   >I think Java is *much* closer to Lisp than to C++.  Don't be misled by the
   >strong typing.  The variable binding model is that of Lisp (in particular,
   >pointers are hidden from the programmer); it's safe; it's GCed.  Because of
   >the strong typing, polymorphism is more restricted than in Lisp, but it's
   >entirely adequate for writing real programs.
   >
   >I guess if C++ and CLOS were your primary exposures to object-oriented
   >programming, you might think that Java was more like C++.  Certainly Java's
   >OOP model is not much like CLOS's.  But there have been a variety of OOP
   >subsystems for Lisp, and Java is relatively close to some of them.

   The most significant criteria for positioning a language, in my view
   is either supports/does not support functional programming well.

   As far as I can tell, Java doesn't support functional programming very
   well, but then again, I havn't really tried.

Java doesn't particularly *encourage* a functional style, but any language
with a GC supports FP much better than any language without it.  (You can fake
GC in C++ using reference counting, but it's a pain.)  

Anyhow, I wouldn't have said FP is what makes Lisp Lisp.  I don't think Lisp's
support for FP is all that great -- Prolog's is better, for instance.

Still, it's true, large Lisp programs usually have lots of little pockets of
functional code, which is not a style you would be likely to find in Java
code, although with sufficient determination you could write it that way.

   >And consider this: implementing Java in Lisp would be far easier than
   >implementing C++ in Lisp.  

   Is there any programming task _not_ easier than implementing C++?

Sure -- Ada :-)

Well, okay, I should have said "... would be easier than implementing *C* in
Lisp".  Since I actually did the latter, I can tell you it's no walk in the
park.

-- Scott