We spent WEEKS getting the permission system right for this project.

This monday 11.30 they tell me we need to change it, preferably by this morning

Now I have 4 branches to merge, I can't remember where the features are, we have regressions all around, the client aren't happy, people can't work, everything explode.

We started this projects 6 months ago, you had all the time to think about it.

Are there any tests? Or is your company one of those bureaucracy first, engineering second' kind of shops?

Always push back. Tell people the risks. Never work overtime, especially if it's their fault things were missed or wrong in planning.

@djsumdog worst, start up. I'm trying to implement tests but i can't write tests if requirement change every 5 minutes, I have actual work to do

@djsumdog I do like overtime, but only if voluntary, not because my manager promised something I didn't agree to. Thinking that way - you should never do overtime if you did do a good job and the estimation wasn't yours. I was once asked where I did come from by some general manager when I entered late (I was younger back then, that's why the person had a attitude towards me and even had the balls to ask). That manager always came ten minutes late en left left ten minutes early while all devs are making 80 hour work week including saturdays and sundays. It was ok, there was a lot of team spirit and we wanted it to happen. The deadline was hard because the date were printed on cards that we were about to send to millions of customers. But so ironic that the guy with such mentality is like "WHERE DO YOU COME FROM" :P He indeed never see's me come or go because in both cases i'm there normally.

In my experience it's always OK to admit that a requirement is overwhelming and will take much longer than a stakeholder imagines.

They will not be upset if you just honestly say you're in the middle of working on 4 other features and this is a big change.

(but they might be short sighted and just say "screw the other 4 features" without realising that'll delay the entire project)

But my advice is: don't say anything in spite like "I told you so" or "We started this projects 6 months ago, you had all the time to think about it." type thing - if you're gonna mention that - talk it over with your PM and mention it a few days later in a retro when you cool off.

Stakeholders are dumb and they hate being reminded of it in snarky ways. But if you can phrase it in a certain nice way they may understand.

@jiraTicket as long you do expectation management right, you can get away with almost anything indeed. Problem with estimations is often that others give a lower estimation than you and that one gets picked and in the end you're the one having to help making it true. Be a teamplayer :P Fuckers :D