RE: Multi-field dynamic array

Jerry:

<snip>
Reminds me of a colleague back in the 70's programming in COBOL on a 
370.  Was having a devil of a time debugging a program; kept getting 
random results even though the data and program remained unchanged.  
Turns out, he said, that a program array was defined with X elements.  
Some variable was using something like X + 1.  Unlike RPG's array index 
error, apparently the COBOL program (at least then) just said, "OK, if 
that's what you want, I'll resolve to that address and give you whatever 
I find."
</snip>

I can vouch for this behavior.  In fact this was the standard hack used if
you needed an array larger than what COBOL allowed.  I remember needing a
very large, multi-dimensioned array with about 150 elements.  The compiler
choked whenever I specified more than 60.  So I left the dimension at 60,
then defined enough "filler" immediately after that to cover the other 90
elements and had no problem referring to all 150 elements!  This is
consistent to what others have been saying about it only "pretending" size
dynamically.

Scary...just scary,
--Bruce Guetzkow