avgagliardo
commented
2 months ago My own two cents on the above:
I would note that that writing a program and running it in a jupyter notebook is not only very easy to do, but it makes the software easier to extend and maintain. I think the best solution would be to write the program as a script, and then present it in a jupyter notebook-- but the difference between the two is not tremendous, we just need to decide on which one our build will be targeted for.
With respect to the entry point of execution, I think there should probably be file called run_bells_inequality.py (or something like that) sitting in the root directory, or just below the root in a directory named /bells_inequality. Any other resources it might need can live in modules nested within the project root and from then on, we would import (ie, in a notebook) or run it directly as the main executable in the project folder.
Finally, I think we should probably save the data (perhaps as json for simplicity, but open to other file standards) so that we can separate the process of running the experiment to obtain the data from actually analyzing that data.
Open to all suggestions!
We need to agree on a project structure.
TO BE DETERMINED
-Is this program meant to be: Run as a jupyter notebook, a python module/program, or both? -Is there a single file that should sit at the "head" or entry point of the execution? -What happens after the program runs. Does the data get discarded? Rendered for visual inspection? Saved to a file for later analysis?
I see a lot of questions about project structure:
33 #75 #63 #61 are the most recent ones I've seen.
We need to consolidate this discussion and come to a consensus or else we will be perpetually stuck in this part of the development cycle.
There are other issues to this point, but tackling this seems like a reasonable first step.