Michael Droettboom

I believe (though I can no longer find the reference) that `python3` is always available if Python 3 is installed, so it should be good enough to just always use...

> Or you want to specifically stick to the approach that Matplotlib uses ? No -- since matplotlib isn't remote in our case, we have a much simpler time. When...

There is some work afoot to make the Jupyter / Binder buttons configurable, so adding pyodide support would be as simple as adding an option to the sphinx conf.py I...

I don't think the docker approach is necessary -- we just need to serve a prebuilt pyodide somewhere (which we already do). This probably will require coordination with the sphinx-gallery...

The wasm backend wouldn't work in a webworker because it requires the DOM. However, running data-only backends in matplotlib seems useful and desirable. Is this also true if you do...

The authors' changes to Numpy are all in this single commit on their fork: https://github.com/fberlakovich/cmq-numpy-ae/commit/301864ac4b85a49b8430ac5fb5382e663f7aabb9 (They admit in the paper that this is not the best way to do it,...

Thanks, I'm excited there is interest to collaborate on this. I think the major gains from this approach would only happen when both CPython and Numpy are participating in adaptive...

> how much brand-new Python C API will be used, and does that do anything to stability or downstream usage of the limited C API I can also imagine that...

> Running a simple python script (python simple.py that contains print("Hello, World!")) currently reads simple.py in full at least 4 times and does over 5 seeks. I have been pulling...

I think it's a good idea in order to make it easier for others. The Faster CPython team's [infrastructure](https://github.com/faster-cpython/benchmarking-public) doesn't actually use `pyperformance compile`, but in the interests of open...