[Info-vax] nice for VMS

[Richard B. Gilbert]

glen herrmannsfeldt wrote:
> Richard B. Gilbert <rgilbert88 at comcast.net> wrote:
> (snip regarding SPARC)
>  
>>> Because unaligned access crashes the program (segmentation fault).
>  
>> I never noticed that one!  I don't do much programming these days.  When 
>> I did, I learned to avoid access violations, segmentation faults, 0C0 
>> ABENDS  (IBM 360/91) etc.  We had some very nice software from Waterloo 
>> Univerity called WATFOR, and, later WATFIV.  It told you exactly where 
>> you had screwed up.  I learned a lot from it about the no-no's of 
>> programming.  It gave a screamingly fast compilation and excellent 
>> diagnostics.  You didn't want to use it for any serious number crunching 
>> however!
> 
> I did some debugging of some number crunching programs
> with WATFIV.  I needed to find some array bounds and undefined
> variable problems that Fortran H wouldn't catch.  The program
> and sample data compiled in about 30 seconds on Fortran H, 
> and also ran in about 30 seconds.  WATFIV compiled it in
> about 3 seconds, but it took much longer to run.  
> 

Yes!  Those were the design goals of WATFOR/WATFIV; a screamingly fast 
compile and good diagnostics!  Princeton ran a free WATFOR/WATFIV batch 
every twenty minutes.  It picked up the really stupid things you did and 
rubbed your nose in them; things like uninitialized variables, and array 
bounds errors.  It was less helpful when you used "+" when you meant 
"-".  I'm not sure if it picked up cruft like variables that were never 
referenced.  The Fortran G and H compilers could give you a cross 
reference listing that helped to clean things up.