Comment on Bell’s Quadratic Quotient Algorithm

Communications of the ACM | , Vol 13(9)

This short note describes a minor inefficiency I noticed in a hash-table algorithm published by James Bell. It got me thinking about hash tables, and I invented what I called the linear quotient algorithm–an algorithm that seems quite obvious in retrospect. While I was running simulations to gather data for a paper on that algorithm, the latest issue of CACM arrived with a paper by Bell and Charles Kaman titled The Linear Quotient Hash Code. I had devised three variants of the algorithm not contained in their article. So, I wrote a paper about those three variants and submitted it to CACM. The editor rejected it, without sending it out for review, saying that it was too small a contribution to merit publication. In the next few years, CACM published two papers by others on the subject, each completely devoted to one of my three variants. (Those papers had been submitted to different editors.) My paper, which was probably a Massachusetts Computer Associates (Compass) technical report, has been lost. (Compass went out of business a few years ago, and I presume that its library was destroyed.) The linear quotient method is probably the most common hash-coding algorithm used today.