I've got a little pet theory related to this. Which is that no one really solves difficult problems. They turn difficult problems into easy problems and then so

I've got a little pet theory related to this. Which is that no one really solves difficult problems. They turn difficult problems into easy problems and then solve that. Often that literally means "We have this class of problems, I'm only going to solve the easy subset". And I've observed repeatedly people say "No we have to solve the real, hard, problem" - and a decade later, with a team of a hundred engineers, they've not really got anywhere. Applied to this, that theory is very simple. It's enormously difficult and expensive to keep old code running, especially when the engineers who originally wrote it have left. Not only is it difficult, but the engineers who have to do maintain it will hate it, churn themselves and you'll be back to square one. So, solve the easy problem. Get your new guys to write something new, and give up on the old software once all the people who knew how it worked quit. And it's easy! No one left at the company is going to argue to keep the code. The new people will enjoy writing something new. Oh and people will be amazed at how happy and productive your team is - because you literally just choose not to do things that are difficult and no one appreciates. But you may ask, what if the product is essential, you can't just abandon it. Well, call me flippant but if it were essential you shouldn't have let the engineers quit. And if you're really going to saddle me with it, I'll quit too. Then you still don't have your software, and it's not my problem - it's yours.