Sorting and Searching
Doghouse – Algorithms and Books

A look at the history of computer memory and a classic algorithm text.
When I started programming in 1969, even mainframe computers from IBM might have had only 1MB of main memory and two or three disk drives that were measured in hundreds of megabytes, with multitrack tape drives that would supplement that storage. Any type of parallelism in programming typically stressed the goal of keeping the data that you needed coming into main memory, while you were also trying to get the processed data out to "stable storage," either to a disk or directly to a magnetic tape.
Some of the first IBM computers I programmed used an operating system called MFT (Multiple Fixed Tasking) which had up to 16 "partitions" in memory, each of a fixed size set when the operating system was configured.
When a job was started by the operators, the Job Control Language (JCL) told the operating system how much main memory it would take, and the job was slotted into a memory slot that was large enough to hold it. If the creator of the JCL specified too much memory, that meant memory was wasted. Too small an amount and the program might not finish, or even start.
[...]
Buy this article as PDF
(incl. VAT)