Open deniz195 opened 5 years ago
I am not really sure about this but what I can say is that the support for this in pint is stalled
My suggestion here is: 1.- Just like we did for pandas, lets create a pint-measurements package 2.- Remove all measurements related things from pint and use pint-measurements (when installed) 3.- Let it evolve independently and with a better defined API and rationale.
Hi, I looked a bit into this. I think removing the Measurement class and going back to using Quantity class with a ufloat magnitude (for the people who want that) should work fine.
Do you want me to prepare a pull request for that? Or do you want to wait for removing this?
The following example shows that the equality operator between two Measurement objects is surprizing (in a bad way):
One can debate the semantics of equality between two values with an errorbar. Still, in this case the objects m1 and m2 are made in the same way from the same registry, and so I would expect that they compare equal.
Currently this is the case for Quantities:
Several possibilities:
.value
and.error
.value
(#preferred)Thoughts?