This is quite reasonable, aside from the variable name which should be isAdmin. A user either is an admin, or isn't. Unless we don't know, then it's null. You are correct this is bad if the point was to represent roles, but it's not supposed to.

[–] orgrinrt@lemmy.world 1 points 1 day ago (6 children)

Admin is a role though, was my point. And besides, if you check for three different states, and you decide to go with a boolean to represent that, I really find it hard to believe anyone would think it reasonable. It’s valid and it’s practical, but can you really say it’s reasonable?

I don’t do typescript, but wouldn’t a union of a null and a bool be just more resource intensive than simply using an unsigned byte-sized integer? I struggle to find reasons to ever go for that over something more reasonable and appropriate for what it attempts to represent (3 distinct states as it stands, and likely in future more than just 3 when they have a need for more granularity, as you’d often do with anything you’d need an admin role distinction in the first place), but likely I’m just not familiar with ts conventions. Happy to hear the reasoning for this though.

[–] shape_warrior_t@programming.dev 1 points 20 hours ago (1 children)

My preferred way of modelling this would probably be something like
role: "admin" | "regular" | "logged-out"
or
type Role = "admin" | "regular";
role: Role | null
depending on whether being logged out is a state on the same level as being a logged-in (non-)admin. In a language like Rust,
enum Role {Admin, Regular}
instead of just using strings.

I wouldn't consider performance here unless it clearly mattered, certainly not enough to use
role: number,
which is just about the least type-safe solution possible. Perhaps
role: typeof ADMIN | typeof REGULAR | typeof LOGGED_OUT
with appropriately defined constants might be okay, though.

Disclaimer: neither a professional programmer nor someone who regularly writes TypeScript as of now.

[–] orgrinrt@lemmy.world 1 points 14 hours ago* (last edited 1 hour ago)

Yeah obviously with constants for the set roles per value. Some languages call them enums, but the point is that what we pass and use is always still the smallest integer type possible. With the extra bonus that if the roles ever become composable, the same value type would likely suffice for a bitflag and only thing needing refactoring would be bitshifting the constants.

But anyway, this turns out to be the weirdest hill I find myself willing to die on.

load more comments (4 replies)