Roger Glover wrote:
> HOWEVER...
> Of the four cases I cited, the seismic data analysis case is
> the only one for which your inclination may be justified,
> but I only say that because I know so little about that
> case. In each of the other cases, inheritance was the key
> desired feature that moved them away from Fortran, although
> exception handling and templates were also valuable in more
> than one case.
Can you be more specific about the importance of inheritence to these codes?
In what I have read about C++ and scientific programming templates have been
emphasized more than any other feature. Which aspect of ineritence was more
important to them, incremental development (implementation inheritence) or
polymorphism (interface inheritnce, how I hate that term)?
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|