So why is estimation so hard?
Software isn't like physical acts of creation, like bricklaying. Physical creation is tactile; the constraints are clear to see. This makes it easy to intuit estimates for bricklaying.
But the constraints on software development are invisible. How long does it take to arrange information and communication structures?
Another MMM thought: execution for software projects is really, really hard.
When the book was written, "execution" of software design and implementation was foetal; I'd say the distribution of people who know how to ship software is multimodal at best.
Estimation techniques conflate effort with progress (ie: the mythical man-month).
Even though estimation is a crap shoot, engineers struggle to push back on estimates.
The feedback loop to zero in on accurate measurements sucks.
An idea that popped into my head reading The Mythical Man Month:
Programming sucks, because it requires impeccable communication and coordination skills, because you make things for other people.
So it's possible that your authority to build $THING doesn't extend to support your responsibility to build $THING. You will have to fight uphill against that to fix other people's things, or get the right info and understanding of other people's ideas, and so on.
https://www.threads.net/t/CuVjWxcJ22Z
“2m curious about Threads, retention rate unclear”
Adam Byatt
Leisa 🇦🇺
Missing The Point