Emil O W Kirkegaard

Over 1k lines of code per commit?? I guess they probably squash their PRs, still pretty wild though

https://twitter.com/GottfriedHerzog/status/1116458117712285697 …

Makes sense, I thought it could be something like that

Could be reality, developing an algo is not the same thing as writing production code.

I didn't do any extensive analysis, hence the question mark. Looking at the commits count shows an obviously less stark contrast (6.3 ratio, vs. 353). I looked over his recent contributions, and they seem pretty normal to me.

I think commits are a fine metric baring explicit gaming tactics. You can do a fine-grained analysis of the commit history if you are not okay with this metric. As for pubs, she has 10 and he has 11, both have 2 as first authors (arXiv data).

I wandered through git blames of the code, looks like 70% that achael guy, followed by chi, then michael, then katie. The involved stuff is all achael and michael, all katies changes are tweaking parameters / configs etc.

I assume she's involved in a more research/scientific role. I think she would need a deep understanding to be making the adjustments she did. But certainly the bulk of the algorithms + implementation work is down to achael (& to a lesser extent michael and chi).

Ah, actually it looks like she also committed some of the early code from a different account. So she could have written that. e.g.: https://github.com/achael/eht-imaging/blame/master/ehtim/imaging/starwarps.py …