ESTIMATING FUCKING HOURS.

Well not literally, that only takes 10 minutes.

But software estimation... anyone pretending to be good at it is a dirty liar. Adding a button? Uh, let's say 2 hours, I mean I gotta poop in between as well, and it's probably some broken bootstrap theme with hacked custom margins.

Building a commenting system coupled to an ancient CRM? Uh... maybe one day? Maybe a month? I don't even know what the CRMs looks like? You won't show me because "that's irrelevant"?

WELL THEN I DON'T FUCKING HAVE A CLUE.

And in the time we spent on discussing time estimations, I could have written like half of the feature... or zero... because you still haven't fucking shown me what the CRM API looks like.

YOU KNOW WHAT I'LL GIVE YOU AN ESTIMATION. ME. VACATION. DONE IN 6 WEEKS.

estimations suck :D

Naw. Estimations are easy. Magic, even. You just think real hard about every line of code you need to write and imagine how long that will take. I mean, when I call a roofing contractor to fix my shingles, he just measures it all out and writes it up! How hard can it really be?

We estimate a time, then double it. This way if it turns out to be a pig then there's time to deal with it and if it's easy then we're great.

@bash-apprentice poker planning I like alot better, comparing past with recent sprint items

@bash-apprentice

Guessing how long it takes to build a REST API with well defined functionality is doable. You double your expectations, and still overshoot by 25%, but... kind of doable.

The big problem is that with my current projects it's all completely unknown territory. There are no external clients involved, it's experimental stuff.

Things like: "Lets write a database query proxy with dynamic masking filters which works with our heavily extended cluster", "We should use AI to mine social media, weather patterns and traffic flow data to predict event attendance" and "We will port our incredibly complex old invoicing system from PHP to Rust to improve reliability and scalability. But nothing about the current system is documented."

So I'm just rolling 10 D20 dice, and fill the sum in as an estimation of days.

I feel ya. The worst part about estimating is that managers and execs are too stupid to understand that it's just a fucking GUESS and for some reason feel the need to hold you accountable for your estimates as though they're some kind of sacred promise, so that if you realize halfway through working on something that it's going to take longer than you originally thought, your only option is to hack some shit solution together to get it done by the deadline.

…and then they wonder where all this technical debt comes from!