wow

["Ji-Yong D. Chung" <virtualcyber@erols.com>]

    Hi,

    I have been looking at some scheme systems, I came across MzScheme..
It is an impressive body of work that you produced.

    I also tested and ran your interpreter, and compared how its execution
speed with other scheme.interpreters.  For example,
your interpreter seems to be 2 to 3 times as fast as MITScheme
to run a set of scheme expressions I devised.  In another instance it is
about twice as
fast as another interpreter (RemoteScheme -- which interfaces with CORBA,
I ran both yours and that interpreter in Windows NT 4.0, on old Pentium 200.
The other interpreter was compiled on MSVC++ 6.0, Release
build.  Your executable was downloaded directly from MZSCHEME website.

    I have a general question ... why is your interpreter behave as it it
were
on speed? Is your interpreter performing optimizations?  Do you think
MzScheme's
performance is attributable to the particular compiler that was used
to generate interpreter executables or rather due to its internal
optimization technique
(if it is optimizing ... that is) I know that RemoteScheme does not perform
s-expression
optimizations, and it has its own internal stack.

    Any comments would be appreciated.

Thanks in advance.