• Background: USENIX started in 1975. In the early 2000s, DTrace was announced at USENIX ATC. But the 2004 ATC was very academic with only PhD students presenting.
• Discussion and Concerns: There was a lot of discussion about the decline of industrial participation. Ted Leung noted this and referred to Rob Pike's polemic. The author wrote about it multiple times, asking questions like "whither USENIX?" and "whither systems research?".
• Analysis: Open source became more important. Leading-edge innovation is in deployed, production, open source systems. USENIX ATC struggled with an overly academic focus and drifted from practitioner relevance.
• Keynote in 2016: The author was asked to give the ATC keynote in 2016 and presented "A Wardrobe for the Emperor", highlighting the error of insisting on conferences as the only publishing vehicle. Rik Farrow wrote about the talk, showing that this was a long-standing concern within USENIX.
• Current View: While ATC died in academia, it wasn't just academia that killed it. Online conferences have their advantages. ATC was a unique forum for presenting pioneering systems work, and the best works still retained this spirit. The last year's Best Paper at USENIX ATC is on a relevant system.