I'm wondering what are good cases for WET (Write Everything Twice (or more)) code. I could kind of see arguments on both sides for this particular case: - ThreadSky

kyleshevl.in • 12 days ago

I'm wondering what are good cases for WET (Write Everything Twice (or more)) code. I could kind of see arguments on both sides for this particular case:

Comments

kyleshevl.in•12 days ago

I am purposely using this example because writing the conditional check _is_ relatively easy.

What are the factors that sway your decision? Where is the line between one way or another for you?

islamnaasani.com•12 days ago

I think i would extract the logic if I find myself updating both repeatedly

or causing a bug because one of them wasn't updated

faithfinder.bsky.social•12 days ago

This particular one isn't about WET for me. It's not really business logic, there's nothing to encapsulate.

kyleshevl.in•12 days ago

Sure, it’s not a core piece of logic; you could have access to this with lodash for example.

faithfinder.bsky.social•12 days ago

Meh, pointless. I don't trust these utility functions, you never know what they handle. It could include empty array/object handling as nullish if the author is Python lover, for example. It just wastes my time with having to check contents/docs

kyleshevl.in•12 days ago

Interesting. In the context of JS/TS, I would posit the concept of nullish is defined in the way it’s used in nullish coalescing.

faithfinder.bsky.social•12 days ago

We have a saying in Russia, roughly translated: "The shed has "Dick" written on it, but there's actually firewood inside" =P I don't trust that the person who wrote the function knew what they were doing. Also, your naming is great, but there's equal chance it would be called `isEmpty`.

kyleshevl.in•12 days ago

I see your point. I see `isEmpty` as a different check, one that would include [] or {}, but fair enough.

I was curious what people would think about this. I've used both patterns, don't really mind either.

volkan.io•11 days ago

A little copying is better than a little dependency. It’s a balance.

slime.camp•12 days ago

Kinda difficult with the example because a nullish check is easy and straightforward. It’s more work to go and find the utility function and import it than to just use a thing == null test

kyleshevl.in•12 days ago

That's exactly why I'm using this example. Where's the line?

Is it number of times this check is made? Does that number change based on the ease of writing the conditional check? Does it change on how discoverable it is?

slime.camp•12 days ago

i mean in this case I think the function is strictly worse., it’s not shorter or faster and it’s enough of an idiom that I think you could say the semantic value of the named function is less than something more complicated

kyleshevl.in•12 days ago

I'll maybe give you semantics. We have "nullish coalescing", so "nullish" should be a known concept.

However, it is shorter, even if it's only a few characters, so I won't budge on that. If your variable names are longer, you save even more.

I'm not saying isNullish is better either, just curious

slime.camp•12 days ago

It’s not shorter when you consider the import, not to mention you can just do thing == null (double equals) and its equivalent to what you wrote

kyleshevl.in•12 days ago

Lot of codebases disable == with an ESLint rule. Haven’t worked in one that allows it in a long time.

slime.camp•12 days ago

I’m writing on my phone during my lunch break so sorry about brevity and being unclear

kyleshevl.in•12 days ago

You’re good. No worries.

Comments

Posting Rules

Reply