project decomposition, and eventually project simplex, are ‘working through’ two landmark books. computer solutions of linear algebraic systems, forsythe and moler, 1967 and advanced linear programming computing techniques, orchard hays, 1968. they’ll be referred to so often going forward, it’s probably best to establish a naming convention early on. las for linear algebraic systems. alp for advanced linear programming.

both are concentrated summaries from first-generation coders, discussing their work from around 1950 through the mid sixties. and both concern code for the ibm machines from the cpc card programmed calculator, through the 701 defense calculator, up to the 7094. roughly speaking, the examples in both concern the 7094. it’s the baseline system assumption for both.

a few other manufacturers are mentioned occasionally, always as brief ‘comparisons’, and notably not including dec. the large scale matrix math of these books necessitated ‘large systems’, and probably the pdp-10 was just beginning to make dec viable in this field. this was an interesting twist of events, as dec was the child of project whirlwind / sage. the 701 defense calculator was also an offspring of whirlwind / sage.

in a nutshell, dec commercialized the interactive and ‘smaller scale / bottom up’ aspects, while ibm stuck with the batch processed ‘imperial scale / top down’ aspects. eventually dec grew able to challenge ibm directly, with the pdp-10 large scale interactive system. significantly, both lineages retained the thirty-six bit architecture.