Post

it's interesting, like, after ~25 years of making software i'm pretty confident that no amount of deathmarching is going to make the product happen faster but it's only relatively recently that i've settled on a satisfactory theory as to why…

…and the reason why (as far as i'm concerned) is, the actual bottleneck is not lines of code but decisions about what those lines of code ought to say

(in other words it behaves like any other written work, quelle surprise)

0 reposts1 like

To make a poor analogy - it's if the editor and the writer disagreed on what the thing ought to be about and then the writer said "well it's not possible to write those words" but eventually could be talked down to "it's possible but will take 3 quarters" solely because they don't want to do it