Eric Hunsberger

I'm not sure what you guys discussed about `CircularConvolution`, but I would move it to `nengo_spa` as well. As @tbekolay said, there could well be non-vocab versions in there, but...

I think this is fixed in #1535. Here's the plots I get for the first systems:  Here's the plots for the second systems:  Looks fixed to me!

One caveat with this is that many backends will not support this, particularly the neuromorphic ones. That said, e.g. Loihi doesn't support any synapses that can't be made out of...

Here's a link to some timing information: https://docs.scipy.org/doc/numpy/reference/random/performance.html#timings I did a bit of my own timing with the following script. I also found that SFC is the fastest, though not...

Should be fixed by #1387, since that brings in a new more stable matrix exponential and changes how filter coefficients are calculated.

I had always imagined doing this manually, i.e. just going through all the error messages we post and adding a link to the docs to some of them (it wouldn't...

I don't see a problem with allowing `BoolParam` to accept both `bool` and `np.bool_`.

When using part of a dependency (at least in the case of the Scipy stuff I've done so far), there might be a little bit of customization needed (for example,...

One downside to this for a backend like NengoOCL is it means we'll have to either store all the arrays of initial values on the GPU, or copy them over...

Here's some code I wrote previously for a (hopefully) general `TransformParam` class that Ensemble could use. It also has some changes to `Dense.__init__` to allow it to not require part...