>From observation, there seem to be two main uses of OO languages
and OO design in this field:
1) To write a better applications (toolkits, PSEs, ...), front-end to back-end.
2) To provide better reusable tools.
The issues are quite different, from which language to use, to
what objects are defined. If I'm using your CFD code, I don't really
care what language it was written in. If I want to use your sparse
matrix package, and can only use it (easily) from the language you decided
was best, I'll be upset.
I have my own opinions (of course). Rather than trying to do everything
individually, we should try to reach a consensus about what the major pieces
of numerical software are, what "objects" they take as input and produce as
output, and what the interface to those objects should be.
I think one only has to try to piece together the tools in Netlib to
get some sense of the problems with the current (non-OO) approach to
tool creation.
Mike Henderson
mhender@watson.ibm.com
http://www.research.ibm.com/nao
This archive was generated by hypermail 2b29 : Wed Feb 20 2002 - 03:20:05 EST