Hi Michael K, MK > Please define VAR parameters for me versus the alternative > in different terms that I can follow better (I haven't been > following this discussion real close). Reasonable demand. I have just made a half-hearted attempt while discussing the Pascal version of the bare-bones VAR parameter example yesterday. Maybe I'll should do better. You may like the explanation in Jensen-Wirth, 1975, "Pascal user Manual..." which is close to what Bob Purves has said. Certainly there are strong correlations in practice: VAR parameter <--> passed by address value parameter <--> passed by parameter value at call time MK > All I know is that Fortran by default passes variables by > address (which what I think you mean by VAR parameters, since > what 1950?) and that FB passes variables by value. I believe > C and Pascal also pass variables by value by default ... Wow! That means that Fortran and QB both have "VAR parameters" *only*, while FB has has "value parameters" *only*. > which is > why Fortran programmers have to be careful interfacing with > those "other" languages. As an import facility builder C ==> FB, you can well evaluate how much conversion hassle would go away if FB had VAR parameters. If I am not mistaken, you raised the VAR parameter problem in this connection a few months back, and the best workaround mentioned was the dot syntax that Alain has just recalled. The same advantage would probably accrue to translating Pascal and C++ to FB, and double that advantage to translating QB and Fortran. Cheers Larry S.