[pseudohadamard]

And another lesson: If it looks like a complete catastrofuck the minute you walk in the door, walk straight back out again. Sometimes the only winning move is not to play.

This isn't arbitrary peanut-gallerying, I've walked away from several jobs that weren't anywhere near as bad as the one in TFA because I couldn't see any way forward that wasn't going to end in disaster. One of them at least, I'm fairly certain, would have ended up in criminal prosecution.

[normie3000]

What kind of vibes are you picking up on? As a contractor, "project rescue" is a massive opportunity, so a bit of dysfunction is often a good sign.

[roryirvine]

I'm a consultant rather than a contractor, but the "shitshow" projects are amongst my favourites. They offer the chance to make a clear positive impact, which can be hugely rewarding.

From my pov, green flags for a good "bad" project mostly come from the attitude of leadership - especially if they already recognise there's a problem, and that some form of cultural change may be needed rather than expecting a purely technical solution.

On the other hand, if they seem reluctant to consider root causes, or if there's any sense that they're seeking to cover up problems or shift blame, then those are pretty clear red flags. Quick fixes tend to fall apart, and aren't really satisfying for anyone.

[pseudohadamard]

Hard to put into a list, it was just the equivalent of a code smell. Some random examples:

- Fly out to the initial discussion, talk to their architect, some guy they imported from France in an otherwise entirely non-French company a long way from France (not bashing French folks here but just wondering where they dug this guy up). They've read about scalability and told him to come up with a design for that, which is distributed stateless everything (this was years before microservices and CRUD and whatnot). Except that they don't have the resources to build the stateless distributed everything so they're getting all the downsides of stateless distributed everything without getting the scalability that it's supposed to provide. You know its bad when one of the nontechnical sales guys that you're having a few drinks with after work can give you a breakdown if why it won't work.

- Parachute in for the first date, they've hired a bunch of kids straight out of college who are all hacking away at whatever they've been assigned to. So I'm walking around chatting to all of them, asking what they're doing and how long they think it'll take. Most of the time the answer is "oh, probably about two weeks". At the same time I'm running a mental Gantt chart for all the components and figure out that even if they meet their wildly, ludicrously unrealistic time estimates it's going to take them at least two years to get the product out the door. Some of their two-week timelines I estimate would take an absolute minimum of six months work, but probably a lot longer. Or a few days if they use the existing open-source tool that does the same thing, but they're going to do it better so they don't need to use any existing code.

- Site visit, the office manager screws up the travel arrangements, screws up the accom, the developers clock in at exactly 9am and down tools on the minute at 5pm so it's tumbleweeds in the office at 17:00:01, the mgt. is in a separate set of offices on the other side of the building and don't talk to the devs, over lunch the devs are talking about how much longer they'll stay before they start sending out resumes.

- The possibly-criminal one, I'll use an analogy to avoid identifying the actual job but lets say you've been brought in to develop an automated pen-testing toolkit. Not just a vulnerability scanner but something that then exploits the vulnerability. "You don't think that this could perhaps be misused by the people that commissioned it?". "Oh, we were so enthusiastic about the job that we never even thought about that".

You may have noticed a common theme here, talking to the techies and/or taking them out for lunch/drinks. So if there's one takeaway, it's do that. That's how you'll find out whether it's an organisation you want to get involved with or not.