I've been thinking about this issue quite a bit, having heard the same thing in 1970. Predictions were made that CPU time would be so cheap that using high-level languages would be the only rational thing to do. I believe what caused all of the pundit's predictions to go astray were the facts that the increase in speed of the processors was taken out of context with the speed and capacity of DASD, tape, and primary memories. The tremendous drop in prices per unit of storage led to a huge increase in the amount of data maintained by corporations (and I'm sure there were other factors, including government regulations, etc.), and that storage base is still growing in excess of twenty per cent per year.
This has put huge burdens on DBAs in processing this data, backing up this data, and reorganizing this data. Instead of 27 data bases (typical number back in late 1970's) of 500Mb each, DBAs have 10,000 or more data bases of 2-4Gb each. Sorts, data base utilities, back-up software, and almost every other conceivable type of software was subject to massive amounts of data to be processed. Can you imagine if the critical-path components were written in a high-level language? Ironically, CPU time has become almost the sole metric for charging back to customers, either directly or via software charges for increasing the CPU capacity of the installation. In most shops, charging for DASD and/or tape mounts is a thing of the past. This has led to a significant expenditure of resources to lower CPU time, with the interesting result that well over 90% of the instructions executed in a typical shop have assembler as their source language, with emphasis on the word executed. I, for one, would never have guessed thirty-six years ago I would be focused on CPU time, but it is a critical factor, and we go to great lengths to minimize CPU usage. On the plus side, it brings challenging design criteria to the table. Tom Harper IMS Utilities Development Team NEON Enterprise Software, Inc. Sugar Land, TX David Shein said: Oh, how that rings true. Next January will be the 37th anniversary of the FIRST time someone said to me, "don't waste your time learning assembler, it will be completely obsolete within four or five years." If I had a dollar for every time I've heard that (or a variation of it) since . . . wow. What a pile of money that would be. And with that lead-in, I bet you can guess how I still make my living. :) David ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
