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.

On 5/14/2025 9:54 AM, Larry McVoy wrote: IBM did not give bonuses for publishing when I was there (1975-1989), so if that stopped in 2005, it sounds like an aspect of the Gerstner era. IBM when I was there strongly encouraged/rewarded patents, and formally disclosing lesser innovations. There were also significant cash awards for "Outstanding Innovations/Inventions," and annual corporate recognition events celebrating those. Charlie -- voice: +1.512.784.7526 e-mail: sauer(a)technologists.com fax: +1.512.346.5240 Web: https://technologists.com/sauer/ Facebook/Google/LinkedIn/mas.to: CharlesHSauer

On Wed, May 14, 2025 at 12:16:04PM -0700, Alan Coopersmith wrote: What I was at IBM, they paid $500/engineer, with a max of $2000/team for a patent being filed, with another $500/engineer if the patent was granted. It was quite common for IBM'ers to set up patent brainstorming teams of 4 people each (to max out the team limit), and the patents did't have to be related to what your department was working on. You could be in the IBM Linux Technology Center, and if you had an idea for something that might be novel in virtual reality, or massive multiplayer online games, or ways that AI could be applied in a health care context, it was all fair game. A particular patent team could be potentially responsible for dozens of patent filings, so it could be real money, especially if you were a relatively underpaid level 5 SWE. (Yeah, it wasn't as a big of a deal if you were a band 10 STSM.) It wasn't formally organized by management, but the fact that these teams exited and the resulting patents were of relatively low quality (in my humble opinion), this couldn't have come as a surprise to them. (From a defensive perspective, quantity was far more important than quality, after all. The patent system is so totaly screwed up....) But as far as papers were concerned, an accepted paper at ATC was worth (at best) brownie points at performance review time, at least in my part of the company. Combine that with the extreme difficulty in getting travel expense approval, it's pretty clear what management really thought about publishing being part of the job. Creating presentations for VP's, or spending all day in Fall Plan meetings justifying budget line items down to the SWE-month granularity? (And then after those plans become totally obsolete once the budget was handed down from on high, then there was Spring Re-plan....) Sure that was definitely part of the job. Publishing papers? Not so much. - Ted