10 ms·

Metz says she adds TODOs and comments that it has been duped. It's one of those things that requires thought, and she even says it's an advanced technique. Ho

Metz says she adds TODOs and comments that it has been duped. It's one of those things that requires thought, and she even says it's an advanced technique. How many times is too many? I'm not sure, but I can safely say over 100 is WAY too many. Probably 10 is too many. Heck, if you find yourself updating the same code in four different places over and over and over, it's time to abstract. The idea is to let the code sit and let the abstraction reveal itself if there isn't already an OBVIOUS one. As mentioned by the parent poster, you're looking out for these copies to diverge. If four or five copied codepaths haven't diverged after some time, there's a good chance that just from working on it every day you will have realized the proper way to abstract it. You absolutely do have to be careful. But even so, it's arguable that having to update something in 100 different places is better than updating in one place and having it affect 100 different paths where you only want 99 of them (this is some hyperbole, of course).

tcfhgj 2y ago
How do you monitor all code duplications in the code base? Including ones that have been modified slightly ( such as optimizations, name changes, additional statements in between, etc)
- sodapopcan 2y ago
  Tests. AFAIC, this isn't something that should be long living. If it is only duplicated in a couple of places and remains unchanged for years, that's probably fine too, because ya... no one is touching it. If one place does need to change and tests still pass that should mean that the other one didn't need to change and you've reaped the benefit from not prematurely abstracting. There are a lot of ways it could play out, though. Often the duplication is very local and obvious. I think a lot of people take "duplication is cheaper than the wrong abstraction" WAY more seriously than its intended. It's actionable way of saying "don't abstract early" as the counter that is usually: "But then I'll have duplication and DRY is the law." Like EVERY piece of programming advice, though, it's not universal.

tcfhgj 2y ago

How do you monitor all code duplications in the code base? Including ones that have been modified slightly ( such as optimizations, name changes, additional statements in between, etc)

sodapopcan 2y ago
Tests. AFAIC, this isn't something that should be long living. If it is only duplicated in a couple of places and remains unchanged for years, that's probably fine too, because ya... no one is touching it. If one place does need to change and tests still pass that should mean that the other one didn't need to change and you've reaped the benefit from not prematurely abstracting. There are a lot of ways it could play out, though. Often the duplication is very local and obvious. I think a lot of people take "duplication is cheaper than the wrong abstraction" WAY more seriously than its intended. It's actionable way of saying "don't abstract early" as the counter that is usually: "But then I'll have duplication and DRY is the law." Like EVERY piece of programming advice, though, it's not universal.