On Sun, Jan 09, 2022 at 08:42:14AM -0800, Christopher Barker wrote:
> Perhaps it's worth remembering that this thread spun off one about adding
> syntax to Python because the current syntax isn't capable of easily
> expressing an important type hinting concept (i.e. Callable).
I shall quote the PEP's justification (in part) for the benefit of those
who haven't read it.
[quote PEP 677]
There are a few usability challenges with Callable we can see here:
• It is verbose, particularly for more complex function signatures.
• It relies on two levels of nested brackets, unlike any other generic
type. This can be especially hard to read when some of the type
parameters are themselves generic types.
• The bracket structure is not visually similar to how function
signatures are written.
• It requires an explicit import, unlike many of the other most common
types like list.
Possibly as a result, programmers often fail to write complete Callable
types. Such untyped or partially-typed callable types do not check the
parameter types or return types of the given callable and thus negate
the benefits of static typing.
[/quote]
These are legitimate issues with the Callable type hint. Does that mean
that the existing syntax "isn't capable of easily" expressing callables?
I don't think so.
"Easily" is subjective, and at the risk of undermining the justification
for the PEP, I think it is fairly easy:
# type hint for a function of two parameters, the first takes
# an int, float or complex, the second a bool, and returns
# either a str or bytes.
from typing import Callable
T = Callable[[int|float|complex, bool], str|bytes]
It was harder to write it out in English than to write the type hint.
So its not *difficult*, there's one import and one class with a
subscript needed. The type system is perfectly capable of expressing
that. It is no more hard to read than any other expression of equivalent
complexity:
from aardvark import Surveying
T = Surveying((spam+eggs+cheese, eels), foo*bar)
But can we do better? The PEP authors think we can, and I am inclined to
agree with them.
While it is true that "not everything needs to be a builtin", in the
case of typing, we are evolving to use builtins where possible:
* we use int, and always have, not typing.Int;
* what was once typing.List[int] is now list[int];
* what was once Union[float, int] is now float|int
etc, but Callable is the last remaining common compound type without a
syntactic shortcut. Its not that declaring callables are *too hard* (let
alone impossible) but that we can do better than what we've got.
Callable[[int|float|complex, bool], str|bytes]
(int|float|complex, bool) -> str|bytes
The second looks like executable pseudo-code for a function declaration.
It avoids the import and the nested square brackets, and it is shorter
without being so terse it becomes cryptic.
The improved syntax looks just like a function signature with the
parameter names excised:
def func(a: int|float|complex, b: bool) -> str|bytes
which I think is a win.
> So arguing that Python is completely readable for type hints is a bit
> off-mark, isn't it?
What does "completely readable" mean?
If you mean that it is impossible to compose a type-hint of such
stupendous complexity and obfuscatory brilliance that even the infamous
RFC-822 email address validator regex appears trivial in comparison?
http://www.ex-parrot.com/~pdw/Mail-RFC822-Address.html
Of course I don't mean that. Type hints are expressions, and expressions
can have arbitrary complexity, including "too damn high".
I argued that type hints are no harder to read than other expressions of
the equivalent complexity. I didn't argue that they can't be abused or
that the syntax cannot be improved.
More to follow...
--
Steve
_______________________________________________
Python-Dev mailing list -- python-dev@python.org
To unsubscribe send an email to python-dev-le...@python.org
https://mail.python.org/mailman3/lists/python-dev.python.org/
Message archived at
https://mail.python.org/archives/list/python-dev@python.org/message/ZCOX33PTCQI3TUZPIO4GLLUPZJDXYIE5/
Code of Conduct: http://python.org/psf/codeofconduct/
