What is the rationale behind the current implementation of default_multivariate_normal_fn?

The current implementation of default_multivariate_normal_fn looks like (excluding doc-strings)

def default_multivariate_normal_fn(dtype, shape, name, trainable, add_variable_fn):
  del name, trainable, add_variable_fn   # unused
  dist = normal_lib.Normal(loc=tf.zeros(shape, dtype), scale=dtype.as_numpy_dtype(1))
  batch_ndims = tf.size(dist.batch_shape_tensor())
  return independent_lib.Independent(dist, reinterpreted_batch_ndims=batch_ndims)

If name, trainable and add_variable_fn are unused, why do you even require them to be passed as arguments? Furthermore, you initialize the scale parameter of the Normal distribution with dtype.as_numpy_dtype(1), which is as weird as it can be.

Why not simply have a method that returns a Normal initialized with 0s as means and 1s as scales, and require only the shape of that distribution?

Also, why does this function even exist, if you already have default_mean_field_normal_fn, which returns a closure def _fn(dtype, shape, name, trainable, add_variable_fn), which thus has the same parameters as default_multivariate_normal_fn and does more or less the same thing (apart from the fact that _fn uses its parameters to define the loc and scale, rather than deleting more than half of them). For consistency, wouldn't it be better to just use default_mean_field_normal_fn to also initialize the prior (by providing default parameters when calling it, or maybe have an alias, but not another function that does the same thing), or am I missing something?

Start with tensorflow_probability/python/layers/util.py at default_multivariate_normal_fn and compare it with default_mean_field_normal_fn. Trace the callers and API expectations for these callbacks, then document whether their separate signatures and behavior are intentional or whether the proposed consolidation is appropriate.