Q: Why are distributions not broadcastable?

[ikrommyd]

It must have been a design choice to make distributions not broadcast their parameters right? Why is that in particular?
I personally find the scipy broadcasting ability on parameters very useful sometimes

from numba_stats import norm as nbnorm
from scipy.stats import norm as spnorm
import numpy as np

x = np.array([0.0, 1.1])
mu = np.array([1.0, 2.0])
sigma = np.array([3.0, 4.0])

print(spnorm.pdf(x, mu, sigma))
print(nbnorm.pdf(x, mu, sigma))

gives

[0.12579441 0.0972427 ]
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
Cell In[1], line 10
      7 sigma = np.array([3.0, 4.0])
      9 print(spnorm.pdf(x, mu, sigma))
---> 10 print(nbnorm.pdf(x, mu, sigma))

File <string>:3, in pdf(x, loc, scale)

File ~/fun/numba-stats/src/numba_stats/_util.py:90, in _wrap.<locals>.outer(first, *rest)
     88 if first.dtype.kind != "f":
     89     first = first.astype(float)
---> 90 return fn(first, *rest).reshape(shape)

File ~/miniforge3/envs/zfit-dev/lib/python3.12/site-packages/numba/core/dispatcher.py:658, in _DispatcherBase._explain_matching_error(self, *args, **kws)
    655 args = [self.typeof_pyval(a) for a in args]
    656 msg = ("No matching definition for argument type(s) %s"
    657        % ', '.join(map(str, args)))
--> 658 raise TypeError(msg)

TypeError: No matching definition for argument type(s) array(float64, 1d, C), array(float64, 1d, C), array(float64, 1d, C)