[burpsnard]

in the 80s, alongside ibm, cray, heimdall, amdal, hitachi, etc etc you had the volume generics - intel. the generic overtook the custom, enough so the revenue and economy of scale won big. it came accompanied with a sci mindset of functional perf, * / ÷ / ^ / sqrt, rather than tps or $/transaction.

The big iron was a product of large-data-volume business problems - payroll, airline reservations, insurance quotes, credit cards, catalog order stats.

But comp-sci mostly put FLOPS ahead of TPS.

hands up who's heard of data flow programming