News for hackers. Stuff that matters.
Re:Argonne National Lab is attempting to replicate LK-99 by heliophobicdude on Saturday July 29, 2023 @12:23AM (#36916887)
Has there been any serious discussion about using GitHub for peer reviews? Specifically the pull request functionality?
Re:Argonne National Lab is attempting to replicate LK-99 by rsfern on Saturday July 29, 2023 @12:40AM (#36917007)
JOSS does something like this and it’s awesome.

I’m not sure about the pull request workflow though, how would that work for review? Is your concept more for collaboration?

https://joss.theoj.org/

Re:Argonne National Lab is attempting to replicate LK-99 by godelski on Saturday July 29, 2023 @12:55AM (#36917085)
Yeah, but more in the CS world and specifically with ML. There are still plenty of CS groups that don't make their source available.
Re:Argonne National Lab is attempting to replicate LK-99 by einpoklum on Saturday July 29, 2023 @03:35AM (#36917824)
GitHub would not be relevant in this respect because:

* It's owned by a (single) commercial corporation, Microsoft.

* There is censorship both by content and in some respects by country of origin.

* The code is closed.

but otherwise it's an interesting idea.

The C++ standardization committee uses GitHub to track papers submitted to them, see:

https://github.com/cplusplus/papers

Re:Argonne National Lab is attempting to replicate LK-99 by tensor on Saturday July 29, 2023 @09:06AM (#36919922)
I don't understand how pull requests would help with reviews. Peer review is generally looking at the methodology and analysis, not simple typos or wording that might be corrected by a pull request.