Ask HN: Inherited the worst code and tech team I have ever seen. How to fix it?

A little over 2 years I came into a very similar situation, however in my situation we had none of the original developers working on it anymore.

At least from a technical perspective, the key to making this manageable is not a re-write, that’s probably the worst approach especially if you have little to no buy-in from above. From a business perspective, a re-write provides little to no benefit and will only be a large cost and time sink, so you will never get buy-in on that anyways.

The key here is slow, progressive improvement. For example, get it in source control, that’s a relatively simple task and provides an endless amount of benefit. The next step which is a bit more complicated, is get a way to run this in a local development environment.

Getting a local environment for this type of situation can certainly be tough, and you have to be prepared to accept what may be considered a “non-optimal” solution for it. Does your code have a bunch of hard coded credentials and URLs that you would accidentally hit 3rd party services or databases from local and cause problems? The answer to that is NOT to try and extract all those things, because that will take a ton of time and you have no test environment. Instead cut the container off from all internet access and add-in a proxy container to it and give it a container it can proxy through for outbound, then you can explicitly control what it can reach and what it can’t, now you can progressively fix the hard coded problems.

Basically the key is to accept that shooting for “ideal” here is a mistake, and you have to sneak the small progressive improvements alongside meeting the business goals that have been set for the team.

In my experience, if you can sneak some simpler but very impactful changes in, then demonstrate how those help deliver on things, it will be easier to get buy in. If you can point to be being able to deliver a feature weeks ahead of previous estimates and attribute it to say having a sane deployment strategy, or a local dev environment, the advantages become clearer from a business perspective. If you say “we need time to fix this” but have no data or concrete examples of how this helps the business, you won’t get buy in.

In my experience working for other people it normally only gets worse, not better. I don’t mean to say that you shouldn’t put in effort or try hard at work—what I mean is that broken or toxic environments tend to only get worse. People with the ability and the self-respect tend to leave if they see an interpersonal or a structural dumpster fire. I would find another job and I wouldn’t worry about your résumé having a short blip on it. You’d be surprised how many people have a similar story about starting at a company and immediately realizing it was a mistake.

> OP does not have the corporate experience to handle this situation

Even before that, s/he doesn't even seem to understand any measure of business. $20m/year with 3 people is BIG. Any disturbance to whatever makes that happen will hurt the business greatly. When a full rewrite shakes the boat and causes a few millions of revenue loss or the loss of potential market share or opportunities, they will rightly fire him.

Best route would be to improve things without hurting anything. It doesn't matter things are 'properly done'. It matters whether the organization survives and prospers. What's 'proper' is redefined every 3-5 years, most of the time without objective basis anyway. So no need to hurt the business to force the current software paradigm.

> a challenge to secretly improve a codebase without anybody noticing

That is how it should be done in any case anyway. Improvements should get slowly rolled out without disturbing the users and the business.

If such a team of 3 people comprised of script kiddies and 5y PHP coders are going to create a $20m/year product, you can be sure that they will take precedence over anyone who was 'properly' educated in cs when it comes to hiring.

> I'm sure it happens

Yeah it does happen. While using the Internet, quite frequently, you are looking at such products developed by such teams, making millions of dollars a year. Even as the good engineering that is being done at FAANG is now being questioned over profitability, with even Google talking about 'inefficiency'.