At the time of Mythical Man Month, serious code was written in assembler and the though was 3GL and 4GL languages would solve the problem of assembler, but also the overhead around assembler, which it wouldn't.
I think the take away should have been that advances in social technologies, (jira, scrum, in big orgs... VC startup system, stacked ranking, literal religious cult like practices in Amazon*) those technologies are where the big returns are.
`*` By religious cult-like practices, I mean their principles that they have to quote & work into their eval process, and imposed culture, the squeezing out of personal time, sleep deprivation by the pager system. This reminds me of monastic lifestyle, which has a lot of parallels- constantly working, sleep deprivation, culture imposed by rules rather than organically emerging, etc. It's a technology & not for no reason did a lot of monasteries behave like commercial enterprises.
♡ 0 ↻ 0I mean, why not both? Throw unit testing, build servers and LLMs at accidental software complexity, throw better process management and issue trackers and business models (startups+VC+big tech acquisition system) and essential complexity.
♡ 0 ↻ 0MMM seems like a meditation on miscalibration. Were better languages 0.1x better? 2x better? 3x better?
People really like to call out miscalibration. (for example, people get upset about arrogance, is everyone identical or is that person evaluating themselves as 1.1x better? 2x better?)
We'd be better served by doing a Fermi Estimate and stating our confidence intervals.
♡ 0 ↻ 0In that vein, a new technology being estimated to be 1x, 10x, 100x better, is not that bad of an estimate, about 2 degrees of magnitude interval! Better than our range for the count of aliens, where we got a confidence interval of 0-10 degrees of magnitude
♡ 0 ↻ 0