this is very much a nitpick, but I wouldn't call throwing an exception in the constructor a good use of static typing. sure, it's using a separate type, but the guarantees are enforced at runtime
I wouldn't call it a good use of static typing, but I'd call it a good use of object-oriented programming.
This is one of the really key ideas behind OOP that tends to get overlooked. A constructor's job is to produce a semantically valid instance of a class. You do the validation during construction so that the rest of the codebase can safely assume that if it can get its hands on a Foo, it's a valid Foo.
Given that the compiler can't enforce that users only enter valid data at compile time, the next best thing is enforcing that when they do enter invalid data, the program won't produce an `Email` object from it, and thus all `Email` objects and their contents can be assumed to be valid.
This is all pretty language-specific and I think people may end up talking past each other.
Like, my preferred alternative is not "return an invalid Email object" but "return a sum type representing either an Email or an Error", because I like languages with sum types and pattern matching and all the cultural aspects those tend to imply.
But if you are writing Python or Java, that might look like "throw an exception in the constructor". And that is still better than "return an Email that isn't actually an email".
If you have onerous validation on the constructor, you will run into extremely obvious problems during testing. You just want a jungle, but you also need the ape and the banana.
What big external dependencies do you need for a parser?
`String -> Result<Email, Error>` shouldn't need any other parameters?
But you should ideally still have some simple field-wise constructor (whatever that means, it's language-dependent) anyways, the function from String would delegate to that after either extracting all of the necessary components or returning/throwing an error.
munificent|19 days ago
This is one of the really key ideas behind OOP that tends to get overlooked. A constructor's job is to produce a semantically valid instance of a class. You do the validation during construction so that the rest of the codebase can safely assume that if it can get its hands on a Foo, it's a valid Foo.
zanecodes|19 days ago
mh2266|19 days ago
Like, my preferred alternative is not "return an invalid Email object" but "return a sum type representing either an Email or an Error", because I like languages with sum types and pattern matching and all the cultural aspects those tend to imply.
But if you are writing Python or Java, that might look like "throw an exception in the constructor". And that is still better than "return an Email that isn't actually an email".
imtringued|19 days ago
If you have onerous validation on the constructor, you will run into extremely obvious problems during testing. You just want a jungle, but you also need the ape and the banana.
mh2266|19 days ago
`String -> Result<Email, Error>` shouldn't need any other parameters?
But you should ideally still have some simple field-wise constructor (whatever that means, it's language-dependent) anyways, the function from String would delegate to that after either extracting all of the necessary components or returning/throwing an error.