[TUHS] Re: ATC'25 is the last ATC?

On Tue, May 06, 2025 at 07:55:26PM -0700, Larry McVoy wrote: I recently came across Bryan Cantrill's relections on the end of Usenix ATC[1], which is well worth the read. It references a number of also musings[2][3][4][5] that are also (in my opinion) very thoughtful, and shows that the tension between practical industry work and academic work is a very old one. [1] https://bcantrill.dtrace.org/2025/05/11/rip-usenix-atc/ [2] https://bcantrill.dtrace.org/2004/07/06/whither-usenix/ [3] https://web.archive.org/web/20081006150917/https://www.allthingsdistributed… [4] https://bcantrill.dtrace.org/2004/07/08/whither-usenix-part-ii/ [5] https://www.usenix.org/system/files/login/articles/login_fall16_01_farrow.p… One of the interesting observations in [2] was that the program committee for the 2005 Freenix / Open Source program committee was solely from industry, and since I was one of the folks who participated in Freenix Program Committees (although I wasn't on the 2005 Freenix PC), and Bryan's concern that it would be the "dumping ground" for subpar work that wouldn't be accepted at the "big boys table". It's a lot more than that, though. I was at IBM at that time, and at IBM, right about then the bonus for getting for getting published at a conference had gotten eliminated. It was very clear that as far as IBM management was concerned, conference publication didn't matter. If you filed a patent, you would get paid a cash bonus. If you submit to any conference --- you wouldn't. On top of that, writing a good paper that would be accepted at conference where you are competing with academics working full-time trying to get tenure, is a lot of hard work and takes a lot of time. So speaking as a working engineer, you'd have to be especially passionate about writing a paper (perhaps because you came from the academic background), to think it was worth it. Open source made this worse, because if you're not going to get paid by your employer, and won't be given work time to work on a paper, if you're going to have to contribute on your own time, the question then becomes, which is better for your future employability? Publishing a paper, or contributing to an open source project, and then becoming a major contributor to an open source project? Finally, what industry conferences run by organizations like the Linux Foundation optimized for the value that most industry players cared about, which is the "hallway track". And that's ultimately what doomed Freenix. Freenix still tried to make people write papers, in the hopes that they could hone their paper-writing skills so they could eventually "graduate" to Usenix. But engineers didn't care about writing papers, and if you could just submit a talk proposal to the Linux Plumbers Conference, which do you think they would choose? If the main reason why you attend, and why your manager would approve your travel expenses, is because you can get face time for the people who will approve your open source contributions (which is what the managers care about), the choice between Freenix and the Ottowa Linux Symposium or Linux Plumbers wasn't even close. All of that being said, it is sad that Usenix ATC is ending. I will admit; I haven't attended in many years, mostly because of time and travel expense constraints. And from a personal perspective, last year, I spent some of my own money (with hotel costs generously covered by the Linux Foundation) to travel and present[6] at the Linux Foundation Members Summit. But could I say that I would spend my own money to attend a Usenix conference? Well, I never have, so probably not. :-/ [6] https://docs.google.com/presentation/d/11rMc8PBeyMItV6hv31OHSZ626_6FCZxjX6Z… - Ted P.S. Admittedly, the presentation at the LF Members Summit wasn't a technical topic, and the whole point was for me to make a point to executives at companies which use Open Source.