> This breaks down in the etc. What if an appointment is not valid on them weekend, or a holiday, or when someone's child has soccer practice, or when that appointment conflicts with another one.
Either the rules are an instrinsic part of an appointment or they're not. If you need something that represents exceptions to the state of a basic appointment then that's the job of the containing type (say a type called WeeklySchedule or something like that). Just like `Day` can be [1..31], but when used in a a `Date` then the constructor of `Date` wouldn't allow a `Day` of 31 and a `Month` of 2.
> At compile time 99% of the time they are closed. Languages also add features to let you make it closed.
You can't close an interface.
> This is impossible in the real world where hardware issues are real.
Exceptions in imperative-land tend to be used for all errors, not just exceptional ones, but expected ones. The problem is that no one knows what these side-effects are from the surface (i.e. a function prototype doesn't expose it).
Truly exceptional events should probably shut down the application (things like out-of-memory), or, as with Erlang, escalate to a supervision node that would restart the service.
Exceptions exist in functional languages too. However, expected errors, like 'file not found' and other application level 'could happen' errors should be caught at the point they are raised and represented in the type-system.
For example, a common pure pattern would be `Either l r = Left l | Right r` where the result is either a success (Right) or an alternative value (Left - usually used to mean 'failed').
When your function has this in its declaration, you are then aware of possible alternative paths your code could go down. This isn't known with exceptions.
For example the `parseInt` from C# and Haskell (below), one declares its side-effect upfront, then other doesn't. It may be semi-obvious that the C# one must throw an exception, but that isn't the only outcome - it could return `0`.
int parseInt(string x);
parseInt :: String -> Either Error Int
This confusion of what happens inside large blocks of code is exactly the issue I'm trying to highlight.
> Total functions also don't allow for infinite loops which is common place in real systems that mare expected to potentially run forever instead of only being able to serve 1000 requests before exiting.
In pure maths, sure, in reality yeah they do. Haskell has ⊥ [1] in its type-system for exactly this reason.
Anyway, you seem hellbent on shooting down this approach rather than engaging in inquisative discussion about something you don't know or understand. So, this will be my last reply in this thread. Feel free to have the last word as I'm not wasting more time explaining something that is literally about soundness in code and why it's valuable. If you don't get it now, you never will.
Either the rules are an instrinsic part of an appointment or they're not. If you need something that represents exceptions to the state of a basic appointment then that's the job of the containing type (say a type called WeeklySchedule or something like that). Just like `Day` can be [1..31], but when used in a a `Date` then the constructor of `Date` wouldn't allow a `Day` of 31 and a `Month` of 2.
> At compile time 99% of the time they are closed. Languages also add features to let you make it closed.
You can't close an interface.
> This is impossible in the real world where hardware issues are real.
Exceptions in imperative-land tend to be used for all errors, not just exceptional ones, but expected ones. The problem is that no one knows what these side-effects are from the surface (i.e. a function prototype doesn't expose it).
Truly exceptional events should probably shut down the application (things like out-of-memory), or, as with Erlang, escalate to a supervision node that would restart the service.
Exceptions exist in functional languages too. However, expected errors, like 'file not found' and other application level 'could happen' errors should be caught at the point they are raised and represented in the type-system.
For example, a common pure pattern would be `Either l r = Left l | Right r` where the result is either a success (Right) or an alternative value (Left - usually used to mean 'failed').
When your function has this in its declaration, you are then aware of possible alternative paths your code could go down. This isn't known with exceptions.
For example the `parseInt` from C# and Haskell (below), one declares its side-effect upfront, then other doesn't. It may be semi-obvious that the C# one must throw an exception, but that isn't the only outcome - it could return `0`.
This confusion of what happens inside large blocks of code is exactly the issue I'm trying to highlight.> Total functions also don't allow for infinite loops which is common place in real systems that mare expected to potentially run forever instead of only being able to serve 1000 requests before exiting.
In pure maths, sure, in reality yeah they do. Haskell has ⊥ [1] in its type-system for exactly this reason.
Anyway, you seem hellbent on shooting down this approach rather than engaging in inquisative discussion about something you don't know or understand. So, this will be my last reply in this thread. Feel free to have the last word as I'm not wasting more time explaining something that is literally about soundness in code and why it's valuable. If you don't get it now, you never will.
[1] https://wiki.haskell.org/Bottom