Hacker News Reader

Comment by detaro | Hacker News Reader

The fact that they took the feedback last time and decided "lets do more of that" is already a big red flag.

>>>On the Feasibility of Stealthily Introducing Vulnerabilities in Open-Source Software via Hypocrite Commits Qiushi Wu, and Kangjie Lu. To appear in Proceedings of the 42nd IEEE Symposium on Security and Privacy (Oakland'21). Virtual conference, May 2021.

from https://www-users.cs.umn.edu/~kjlu/

If the original research results in a paper and IEEE conference presentation, why not? There's no professional consequences for this conduct, apparently.

Given that this conference hasn't happened yet, there should still be time for the affected people to report the inappropriate conduct to the organizers and possibly get the paper pulled.

FYI .. many ACM conferences are now asking explicitly if an IRB was required, and if so, was it received. This does not prevent researchers from saying IRB doesn't apply, but perhaps it can be caught during peer review.

Btw .. I posted a few times on the thread, and want to acknowledge that researchers are humans, and humans do make mistakes. Thankfully in this case, the direct consequence was time wasted, and this is a teaching moment for all involved. In my humble opinion, the researchers should acknowledge in stronger terms they screwed up, do a post-mortem on how this happened, and everyone (including the researchers) should move on with their lives.

The same group did the same thing last year (that's what the paper is about - may 2021 paper obviously got written/submitted last year), when the preprint got published they got criticized publicly. And now they are doing it again, so its not just a matter of "acknowledge they screwed up".

Given current academia which puts a significant negative on discussing why research failed, I doubt your idea of post-mortems, public or private, will gain any traction.

https://academia.stackexchange.com/questions/732/why-dont-re.... seems to list out reasons why not to do postmortems

There are some venues, e.g. this Asplos workshop: https://nope.pub/

If this is actually presented, someone present should also make the following clear: "As a result of the methods used by the presenters, the entire University of Minnesota system has been banned from the kernel development process and the kernel developers have had to waste time going back and re-evaluating all past submissions from the university system. The kernel team would also like to advise other open-source projects to carefully review all UMN submissions in case these professors have simply moved on to other projects."

I just wanted to highlight that S&P/Oakland is one of the top 3 or 4 security conferences in the security community in academia. This is a prestigious venue lending its credibility to this paper.

I would go even further and say that Oakland is the most prestigious security conference. That this kind of work was accepted is fairly baffling to me, since I'd expect both ethical concerns and also concerns about the "duh" factor.

I'm a little salty because I personally had two papers rejected by Oakland on the primary concern that their conclusions were too obvious already. I'd expect everybody to already believe that it wouldn't be too hard to sneak vulns into OSS patches.

The guy is still putting the blame on the Kernel project for not having a "don't submit bugs" clause.[1]

And insists it was not human research. [1]

How can this type of people be professors?

[1] https://www-users.cs.umn.edu/~kjlu/papers/clarifications-hc....

Reply on news.ycombinator.com