ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe) (12/10/90)
In article <8252@lanl.gov>, jlg@lanl.gov (Jim Giles) writes: > This discussion of pointer-to-pointer arrays vs. 'flat' arrays is > getting out of hand. Extravagant statements about the size of > 'practical' computations vs. 'interesting' computations are not > relevant to the issue. I've changed the title, because when we talk about "pointers" in the same breath as "C" there is a risk of assuming that pointers have to bring in aliasing. It's been a while since I did much number crunching, and it was never as serious many of the posters in this group have to do daily. But I _did_ have a use for triangular "arrays" and for arrays that were "almost" triangular but had one or two "spikes". I had better things to do with the computer's memory than to allocate rectangular regions just so the compiler could do "addressing polynomials". Sometimes, even in Fortran, it would be nice to declare something as of type "symmetric-sum-of-squares-and-products-table-with-margins". Oh well, maybe there'll be an F90 compiler for this machine someday. -- The Marxists have merely _interpreted_ Marxism in various ways; the point, however, is to _change_ it. -- R. Hochhuth.