Excerpts from Jerzy Karczmarczuk's message of 2018-08-06 19:31:50 +0200:
> Why "not nice"? There is no [assembly] code generated for such zombies,
> so why the typechecker should waste time? I think that - at least
> "philosophically" (sorry for the abuse of the word) this is coherent
> with laziness. In these contexts the execution errors won't manifest
> themselves, so why the type errors should?

Hi Jerzy,

I understand your reasoning, and from the perspective of the compiler, I
agree. From the perspective of the programmer, there are two reasons why
I personally would like to see those type errors.

Firstly, when developing source code, I often write helper functions
before using them. Code does not suddenly exist in its finished form, but
there is a process of creating it. Type errors for dead code are useful
in the process.

Secondly, it would be consistent with top-level function definitions.
Those are type checked, even if they are not used.

Either way is fine, the programmer just needs to be aware of it.

Best regards,
Markus
_______________________________________________
clean-list mailing list
[email protected]
https://mailman.science.ru.nl/mailman/listinfo/clean-list

Reply via email to