Methods of Randomization of Large Files with High Volatility
Patrick C. Mitchell, Thomas K. Burgess · Information Technology and Libraries · 1970
Key-to-address conversion algorithms which have been used for a large, direct access file are compared with respect to record density and access time. Cumulative distribution functions are plotted to demonstrate the distribution of addresses generated by each method. The long-standing practice of counting address collisions is shown to be less valuable in fudging algorithm effectiveness than considering the maximum number of contiguously occupied file locations.