I gave you one instruction: don't share your password. We control the server, we can have as many users as we want. Opsec is hard when the users are stupid.

The Slack API will on error return a http 200 with "ok": false in the body.

The frontend rightfully can't call APIs marked experimental from our public API. So teams go and implement everything as private APIs so no one can do anything with our public API since there are hardly any endpoints.

Sure make helm charts depend on other charts that are only accessible inside the production k8s cluster, who would want to have a local deployment of what they are working on anyway

What kind of stupid ticketing system do you have where I can only send enquiries between 0730-1100 each day?

I once resigned the ui of our banking application on my own initiative without consulting with any clients. I then left that job before our next release. Apparently the product owner got flak because a client measured the header with a ruler pressed to their monitor and found the new header was sightly bigger.

Tests define behaviour, if code isn't tested then there is no way of knowing what it should be doing so it therefore cannot be working correctly.

I hate large companies. People assume how things will work and what direction things will go in, teams build competing services, sales sell things and don't tell engineers. No one talks to anyone else.

Why do people like Go so much? The more I use it the more angry I get, it's like the designers thought people were too dumb for C so made a patronising version ignoring the last 20 years of language design. Fucking null checks everywhere! Just give me a result monad!

A client spent a long time in meetings with our management telling them they can't use Apache Tomcat and the software needs to run on IBM Websphere instead. Took ages before they finally put an engineer on the call who pointed out that we are a python house and we don't have any Java applets.

Fucking notion doesn't let you middle click on links, I've pasted so much crap into documents for the whole company to see

I turned on my PC for the first time and someone was impressed. I knew it would be easy after that, luckily didn't stay there long or my brain would have rotted away.

I like writing code, I'm good at writing code, why do I have to be a manager? I just want to work with people, stop pushing me into middle management.

Set our git forge to reject pushes if the commit hash has a 4 in the first few characters.

You could have done a git blame to see why this code was written like it was, or you can just blindly change it without understanding and break something. Guess which option my coworkers take.

I would rewrite the flagship products from the places I used to work at the way I kept saying they should be built. Then release them all under AGPL.

Would telling someone in a code review that their code wouldn't even work and that a simple test would have caught it of they wrote one be considered rude? Should I just tell them exactly what the problem is and move on or make them figure it out?

I want people to be testing the code they write but I don't want to just become annoying to them by telling them constantly to add tests when there is no immediate payoff that these juniors obviously crave. Similarly I don't want to antagonise them when they are so early in their career.

If some code smells even remotely then nit pick it to death. What may seem harmless now will exponentially grow as precedent is set.

If you write ansible playbooks that aren't idempotent then you deserve a quick meeting with the sharp end of a knife. I've wasted too much of my life dealing with this crap.

Voice over NFC.

Fucking pulse audio crap. Something keeps muting my output device and I can think of no way to debug it. And I can't just switch to a good sound system because of every application only supporting pulse now.

Remember there is only choice if your choice is the same as Lennart Poettering's.

The correct place for docstrings is NOT SLACK!

"We need to do operation X about 50 times a second constantly."

X is a database heavy NP-hard problem with an input size of a few million.

And this new thing has to run on the same already stretched hardware that everything else is on.

I spent so long making python notebooks showing graphs why it wouldn't work, eventually someone built the naive exhaustive search anyway to prove it was slow. They pushed that out to production without testing it under any real load. Needless to say clients were annoyed.

20 minutes on a support call with a client trying to tell them how to quit vi. Lesson learned: charge for support by the hour, especially when you aren't allowed to touch the machine your system is installed on.

Why every time I come up with a nice idea for an electrical engineering project it turns out to be far more complex than I could imagine and impossible with my skill level. I have done hours of research and feel like I may as well not bother.

If you want me to implement something using your api, it would be nice if your api worked.

You want to know why I haven't made any progress on my ticket? Well I've spent all day fixing code other people have wrote and somehow got merged in without even running it. I haven't done anything on my code because the project is in a broken state and won't even start. Oh what's that? Nothing is going to change and we are going to continue to work like this because people are used to this way of working? Great, here is my notice.

Support for Websphere applications.

There are rumours at work that we have to make integrations with SAP soon. I might quit now to save myself from nightmares.

I'm sick of all these offices with plywood furniture