Name: sg39081 Date: 10/23/97


Dear Sir,
Java as language takes care of all branhes of
software development. I feel that the scientific
community was lagging behind, because the language
doesn't support 128 bit precission arithmetic.
For most of our numerical computations, the error
tolerances will be in the order of 1.0E-16 or so.
With 64 bit double precission, the maximum
tolerance acheived is in the order of 1.0E-12.

The BigInteger and BigDecimal classes are very
good. But they support persons working in Number
Theory. A support for 128 bit precission at the
language level would be of great help to the
scientific community.

Thanks and regards
avsrao
======================================================================