"Frank Lesser" <squeak at Lesser-Software.com> wrote:
	 ok, I should agree that obfuscation can be solved - but I don't know any
	ST-environment which solves it.

Never mind Smalltalk:  Java code can easily be disassembled.
When I worked at Quintus, we put a fair bit of work into ensuring that
the Xerox Quintus Prolog system couldn't easily be disassembled.  But
one of our competitors disassembled the MC68020 code of the UNIX version
instead (and yes, the object also had a smashed symbol table so that
they "couldn't" do that) and tried to publish our implementation techniques
as their own.

A sufficiently determined competitor will even disassemble Intercal or
BrainF**k.  Squeak currently has enough obfuscation through sheer bulk
and paucity of comments; I really doubt whether adding more would make
it a more useful "product".

	But I don't like to focus on decompilation / obfuscation.
	I believe translation is relevant to deployment - if you can drag out your
	application out of the development environment
	- translate it to Java / C, C++, C# or binary - Smalltalk would be more
	attractive.
	
Why?  There are a couple of Lisp to C compilers; they didn't make Lisp
more attractive.  (I except things like (A)KCL and Gambit, which use C
as a portable assembler, not as a delivery vehicle.)  There's at least
two Prolog to C compilers; I don't see that making Prolog take off.
We _have_ a Smalltalk in Java already (sad really; trying to make fast	
programs in Java is like trying to run with a gorilla on each shoulder).
The Squeak development environment is admittedly large, but there are
3rd-party Java _applications_ that thrash on this 128MB machine.  (Which
is why I now do that work in interpreted Scheme.  Goes faster, too.)