jaw@ames-lm.UUCP (James A. Woods) (02/15/84)
# He confided this to me at the January Uniforum conference. We were discussing chess programming, which is really the motivation for much of his UNIX work. For a veiled reference to this access method, consult the November, 1982 Bell System Technical Journal, part 2, papers on database work at Bell Labs. Thompson's independently developed scheme, not mentioned by name, was compared with a B-tree method as well as another dynamic hashing algorithm. What I could use is a way of employing this external hashing buddy system for full-text searching applications. Multiple (and duplicate) keys would need to be accommodated here. The method of choice so far seems to be inverted files (c.f. Lesk's 'hunt' and 'inv' code). Retrieving arbitrary information in two disk seeks is something to behold. -- James A. Woods (hao!ames-lm!jaw)