[Fwd: speculation about java math methods]

[Ron Boisvert <boisvert@nist.gov>]

-------- Original Message --------
Subject: 	speculation about java math methods
Date: 	Thu, 1 Nov 2007 17:53:00 -0400
From: 	Serge Abramov <s_abramov@rambler.ru>
Reply-To: 	s_abramov@rambler.ru
To: 	boisvert@nist.gov



Dear sirs
After about 3 years of exploration I have come to a conclusion that Java 
doesn't seem to be a best choice as to mathematical computation. I have 
created some programs simulating transient processes taking place in a 
simple autonomous electric power system. They do work but the simulation 
process takes a huge amount of time. I tried to use Jama classes but 
Java virtual machine slowed down simulation all the same. My old Turbo 
Pascal and C programs work much faster. To my mind the only reasonable 
way to employ step integration methods in the area of mathematical 
simulation can be reached with the aid of native methods. Fortunately I 
managed to adjust Netbeans 5.5.1 with C++ module and MinGW C++ compiler 
so my research work can go on. I am going to place these assumption on 
my web site in the nearest future. So if anyone wants to use or create a 
standalone transients simulation program Java can hardy be recommended.
Sincerely
Serge Abramov
St. Petersburg, Russia