[pandler]

> Nobody has any idea, they're just making it up as they go along.

Ha! I came here to say this, and I'm glad you share my opinion. I realize that you're speaking specifically about software estimates, so please excuse me while I take your statement out of context and generalize it to most everything that I've experienced working in software.

I'm biased in my views on this at this point, but I tend to notice people falling into two camps (on a spectrum of course): those who are most comfortable with well-defined tasks and low uncertainty, and those who are comfortable making stuff up and running with it. That's a rough, one-dimensional reduction of many qualities, including creativity, certain types of critical thinking, pragmatism, "self-starter"-ness, enthusiasm, etc. In equilibrium, it's the people who are able to distill uncertainty into something resembling certainty that ultimately provide tasks for the people less comfortable with unknowns.

Corollary to that is the idea that "correct" is not binary but rather a measure of effectiveness by any number of shifting metrics (e.g. performance, readability/maintainability, "correctness", time and money cost, defect tolerance; all under the general umbrella of priorities).

It's one of the things I try to impress most upon junior developers that I mentor, especially when they are just out of bootcamp and haven't learned yet that most problems aren't well defined and don't have an answer in the back of the book, if you would. Essentially, our job is to "make stuff up": always "how", and often "what".

[WJW]

I don't think that it's actually true though. There are definitely areas where I don't have to make it up (anymore, at least). As a trivial example, implementing yet another API endpoint for a web form or something is not something where I have to experiment a lot (anymore).

It's also important to realize that in most fields, people are definitely not just making it up as they go along. A dentist or an airline pilot has extensive training, knows what they are doing and why. This is even true when it comes to estimation tasks.

[daze42]

This seems accurate to me. Maybe the parent is speaking from a place in the Valley of Despair on the Dunning-Kruger curve.

The good news is, they're just passing through.

[jaggederest]

I'm not actually speaking about software estimates there, I should have used asterisks or something. I mean it even more broadly - most people have no idea what they're doing in any given context, let alone in software with all the complexity. The whole world is just winging it, as a general rule.