Charles Stevenson goes on: [snipping the bit about tolerance] >Do you know why dimensioned arrays were first limited to only 1 or 2 >dimensions in Pick or Info-Basic?
There was probably some concern about memory consumption way back when MD and Pick ran with only 64KB of core. I think it was just simpler for the designers to set a low limit. It looks like the subsequent designers never anticipated adding more dimensions, rather they focused on backward compatibility (which in this case has multiple interpretations). >Would it be a terrible complication to the compiler to allow as many >dimensions as the programmer needs? There are two complications: modifying the compiler and modifying the run machine. Modifying any of the compilers to support additional dimensionality would be easy. Modifying the existing op-code structures, descriptors, and run machines are likely a huge challenge these days. This type of extensibility needs to be thought of in the early development stages. Best regards, Gyle P.S. I designed URMA with support for 2^31 dimensions. Yeah, I know, after about 9 dimensions, you probably run out of memory today. Who knows, maybe terabytes of memory will be cheap tomorrow. The motto was, do not limit the design to today's technology. ------- u2-users mailing list u2-users@listserver.u2ug.org To unsubscribe please visit http://listserver.u2ug.org/
