GNU and MzScheme's Internal design

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

Hi, you wrote:

> Too bad you hadn't heard of Fortran.  Of all the languages I have
> seen, it seems to be the one easiest to compile for serious speed.

    I needed something on a server to respond reasonably fast
to a scripting request (XML processing request) from a client 
application.  I suppose I could have implemented a Fortran interpreter
(just a compiler would not have worked) for this --
but I have not seen a working one, and it would have been 
pain in the neck for me to find or write such a thing.
On the other hand, Scheme interpreter was easy to write.

    I would have loved to use MzScheme (it certainly
performed well enough), but I couldn't, for two reasons:

(1)  GNU licensing scheme -- as I was not sure if the 
my app was to be a  GNU too, I could not really take that
chance.  

(2) insufficient number of easily accessible 
documentation of its internal designs and ideas
(I am not putting MzScheme down, as this is
standard practice with most open source libs)

    When one incorporates other people's library,
unless that library is well understood, the borrower
ends up treating that library like a plutonium, because one cannot
modify the borrowed source code with confidence,
for the fear of introducing new bugs and problems.
    
    If I were to use MzScheme for a serious programming
project, I need to understand its subcomponents 
thoroughly -- that basically means I have to single 
step through most of its functions and understand exactly
what it is doing.  That could take a while without
somewhat detailed design documentation -- 
which I could not find.

    It is actually easier to just write a simple Scheme
interpreter that approximates MzScheme's performance.