Just a couple of years ago I would probably be sitting on the camp saying this is too crazy of an approach to work, but glad to have been proven wrong by @PathogenDavid.
I now deeply believe this has the potential to revolutionize the way the .NET community does native interop.
unavailable
Thanks @peter_finnie it was indeed incredible fun, also my first time patching, thanks everyone for making it so enjoyable!
What I feel is missing are more systematic attempts to articulate coherent frameworks of thought. Not sure about the best format is to encourage this, perhaps through focused collections of articles or even books?
This is exactly what we would like to do for the @bonsai_rx IDE. Its a visual programming language, so depends heavily on its graphical interface for editing and debugging. Its basically a micro-VS on top of Winforms with a property grid, a nuget package manager, and debugger.
Systems for pooling grants and redirecting resources with cross-lab agreements would be key to this. In fact examples of this already exist, e.g. see IBL: internationalbrainlab.com/
Mostly we want to discourage wasting money on equipment which will never be used because “grant money needs to be spent” and encourage sharing and reuse of expensive equipment and facilities. Really need to run a 2photon experiment as a control? Engage a lab with a working setup.
If their experiments cost less, then yes. In my view, UBI for science would work a lot as a discussion around the true costs of experimentation. UBI for salaries first, then an open system for submitting experiments and sharing resources.
Interesting that Rich Sutton actually tones it down to “the first” rather than “the” and says it would be good to have others. Of course, the claim to “first” is also problematic itself given cybernetics. Not sure if control theory is itself the first, but definitely precedes RL.
Also very timely and serendipitously very relevant in the context of the above, a great piece by @CyrilPedia: ft.com/content/cda27366-7de5…
Blog post and then tweet with link to the post? Tweets are too ephemeral for historical records of any kind, imo.
The full quote from D. Knuth is: “We should forget about small efficiencies, say about 97% of the time: premature optimization is the root of all evil. Yet we should not pass up our opportunities in that critical 3%.” - I don’t know why people truncate it.
Likewise, thanks for bringing us together! If it weren’t for the DLC Hackathon not sure it would have been done in this crazy year. It was such great fun to brainstorm and make this integration with everyone. Definitely one of the highlights of the year!
We’ve recently done some experimentation setting up @BonVision_ as a VR environment jukebox controlled through python via the python-osc package. It was incredibly smooth integration. I can try to cook up a demo to show this for the next release :)
I finally bit the bullet to lay down in writing my thoughts on the relations between OSS and science. Thanks @AnneEUrai for your thoughtful provocation. I agree long-author lists solve nothing if we are not ready for comprehensive and systemic change. Here is my full take.
Manual steps such as annotations can be stored as well, reviews are issues, discussions can go in project boards, funding can be both crowd-funded or sponsor-led. Science is a public good too, if OSS has changed the world (including science), so can we. fin/