jec (11/20/82)
I've also experienced a few problems with the IBM 5150, mainly the FORTRAN Compiler. But first, I should point out that I've seen and run a lot of IBM software on RGB monitors and I haven't experienced the same resolution troubles. It might be that the resolution of the RGB monitors are sufficiently better, or that the RGB part of the color card is different enough that the problems are taken care of at that level -- I don't know. But as far as the FORTRAN Compiler goes, I have three gripes: (1) RETURN statements do not clear the heap in fuctions and subroutines. This means a function that gets called often will eat up tons of memory keeping the scratch variables around and may eventually crash. The solution to this one is not to use RETURN and just end your fuctions or subroutines with END. (2) FORMAT(A) does not work as the manual says it should. The following program will bomb-out: CHARACTER*10 MYVAR . . . READ(*,100) MYVAR 100 FORMAT(A) . . . WRITE(*,100) MYVAR If you enter "THISISATEST" ( >10 characters) it will work fine. If you enter "TEST" it will bomb-out with some error message that doesn't really apply. (3) IMPLICIT statements behave erratically. I can't pin down the exact problem, but there is one. It tends to be between a main program and a subroutine or function. If anyone has a cure for these, I'd appreciate it. I've written to IBM about it and talk to ComputerLand in Mishawaka, but so far I've heard absolutely nothing from either. I've heard rumor of a FORTRAN update or newsletter from ComputerLand, has anyone heard of it? Mike Conley ...decvax!inuxa!isrnix!jec