Subj : Re: Industry Calls for More Foreign Programmers To : comp.programming From : blmblm Date : Tue Aug 23 2005 06:11 pm In article <1124793064.245219.216620@g43g2000cwa.googlegroups.com>, wrote: > >blmblm@myrealbox.com wrote: >> In article <1124443672.965711.76370@o13g2000cwo.googlegroups.com>, >> wrote: >> >> [ snip ] >> >> > >> >The brutal truth: if, for example, a programmer has specialized in >> >financial Fortran for several years, he will not have adopted the >> >structured, storage management, and (god forbid) object-oriented (if >> >any) improvements to Fortran but will stick to canonical Fortran, >> >including overuse of overexposed common data and Equivalence for shits >> >and giggles. >> >> What do you mean by "canonical Fortran", and what objections if any >> do you have to making use of features included in Fortran standards >> later than FORTRAN 77? >> >> (Really, I suppose I'm asking whether you're aware that Fortran has >> evolved significantly from the days when there was no way to define >> data structures other than arrays, or allocate memory dynamically, >> without using non-standard extensions. Maybe you are.) > >Maybe I am. Oops, I am. However, I have experience, observer and not >participant, of two Fortran communities. One was the Princeton >supercomputer community in the early 1990s, and the other was the >Chicago rocket science financial community in the early 2000s. > >In both, developers AVOIDED the new features of Fortran for no reason >except conservatism. > >In the latter community, using the newer features was frowned upon to a >degree. Actually, given the timeframe there could be some reason other than sheer conservatism -- the way I understand it, the first compilers for Fortran 90 and later standards weren't very good at translating the new features into fast code. (So which side of this argument am I on, anyway?) -- | B. L. Massingill | ObDisclaimer: I don't speak for my employers; they return the favor. .