The ACM.org website published the work of a team at Carnegie Mellon (#CMU) which was said to include source code. Then the code was omitted from the attached ZIP file, which only contained another copy of the paper. I asked the lead researcher (a prof) for the code and was ignored. Also asked the other researchers (apparently students), who also ignored the request. The code would have made it possible to reproduce the research and verify it. ACM ~~also ignored my request and~~ also neglected to fix the misinfo (the claim on the page that source code is available). Correction: ACM replied and tried to find the missing code but then just gave up.

It seems like this should taint the research in some way. Why don’t they want people reproducing the research? If the idea is that scientific research is “peer reviewed” for integrity, it seems like a façade if reviewers don’t have a voice. Or is there some kind of 3rd party who would call this out?

you are viewing a single comment's thread
view the rest of the comments

[–] themoonisacheese@sh.itjust.works 18 points 2 weeks ago (1 children)

I have a contact at cmu's student journal (for what that's worth). They might be interested in this, maybe if they start prodding around the researchers will realize the mistake. What's the publication?

[–] plantteacher@mander.xyz 5 points 2 weeks ago* (last edited 2 weeks ago)

Oh, wow.. I wasn’t expecting that reply. I was actually looking to discuss in general how to address this variety of issue. It was a few years ago but the code would still be interesting to see. I dug this up:

https://dl.acm.org/doi/10.1145/2911988

And now that I dug back into this, I must make a correction. ACM replied to say they are looking for the missing material.. then they never found it and they dropped the ball at that point and also neglected correct the description. AFAIK, ACM did not try to reach the researchers, who ignored my inquiries.

(Irrelevant trivia: ACM used to be in Cloudflare’s access-restricted walled garden, making it difficult to access research. They are still in that shitty place but at least they are now whitelisting Tor which slightly reduces their exclusivity.)