Re^4: End of Native Code?

It is too easy for the scientists and engineers to write their own data reduction software in it.

Are you implying that scientists and engineers shouldn't write their own code (and CFD, numerical relativity, etc are most emphatically not data reduction software; they're models)?

If so, I suggest that CS majors start getting better backgrounds in physics, mathematics, fluid mechanics, and chemistry. Very many of them do not have the basic tools needed to write or maintain these kinds of code, and if a CS graduate can't manage ODE, let alone PDE, they can't even have a sensible conversation about Navier-Stokes, NRQM, or QCD (not that I can have a sensible conversation about the latter two.)

By the way, it's not FORTRAN; it's Fortran. It has been for about 20 years, since X3J3 finished the F90 standard

emc

e^(π√−1) = −1

Comment on Re^4: End of Native Code?

Replies are listed 'Best First'.
Re^5: End of Native Code? by girarde (Hermit) on Jun 29, 2006 at 19:31 UTC
I am not implying that at all. The mathematics in these problems is such that most computer scientists just are not in the game, and really shouldn't have to be: that's what Fortran is for.) And when I used it, it was still FORTRAN. :-)	[reply]
Re^6: End of Native Code? by swampyankee (Parson) on Jun 30, 2006 at 15:54 UTC
When I, too, started it was FORTRAN. Indeed, on the machine upon which I learned to program, it could be nothing else: 6 bit characters (packed 8 characters per 48 bit word, it didn't yet have a character data type). I think they had, in addition to FORTRAN and assembler, COBOL, ALGOL (I'm not sure which flavor; this was in the early 70's), LISP, SNOBOL, ICEBALL (a dialect of SNOBOL), and IITRAN, (there was, iirc, one other installation of IITRAN). emc e^(π√−1) = −1	[reply]