Talks & Interviews
Geometry to Algebra and Back: 4000 Years of Papers
Video of an invited talk that I presented at the 2023 edition of the Papers We Love Conference at Strange Loop. A complete transcript with corrections, references, and additional notes can be found here.
Stop Writing Dead Programs
The Aesthetics of Programming Tools
Clerk is an open source programmer’s assistant for Clojure that combines the advantages of notebooks like Jupyter, Smalltalk-style Moldable Development, and using one’s favorite editor. We published a paper on the system at Programming '23, though I strongly recommend the online version over the print one.
Classic HCI Videos
A collection of classic UX/HCI demo videos produced during the golden age from 1983-2002, presented in the graphical style of the era.
What Does it Mean to Buy a GIF?
An attempt to clarify the NFT (Non-Fungible Token) phenomenon, with an eye toward the history of art markets.
An explanation as to why this site is the way that it is, RSS-feed and all, in the age of social media.
Leopards in the Source Code
An exploration of code re-factoring through the lens of a Franz Kafka parable.
Finding the Best Fighters
An computational method for rating MMA fighters based on their performance in PRIDE FC and the UFC, along with some visualizations of the results.
Past Project Highlights
Made to Measure
Berlin-based art collective Laokoon produced a documentary film in 2021 to highlight the privacy implications of using large social media platforms. I helped with the project by training an ML-based voice clone of the film’s protagonist in two languages. Collaboration with Studio NAND, who produced excellent process documentation about the project.
Anyone Still Working Here?
Startups, Research, and Open Source
We developed a sophisticated search interface for automotive repair information. As part of this, I created a natural language processing system that combined neural networks and classical AI/semantic web techniques. Acquired by Advance Auto Parts in 2013.
We built one of the industry’s first commercial Complex Event Processing systems for streaming data. It was used for fraud detection, trading platforms, and other soft real-time information processing tasks by many giant companies. My paper on the storage system from VLDB 2007, A Log-Structured Store for Streaming Data, is available here. Acquired by Sybase in 2009, now an SAP product.
We built the first wide-area high-availability system for Unix-like computers in 1998, which was used by Apple, Ford, State Street Bank, and numerous other Fortune 500 companies to provide multi-site failover for their services. We also created a highly available, self-healing clustered filesystem, after which we were acquired by Veritas in 2002.
I built a successful B2B ISP/ISV in the early/mid-90s using commodity hardware and a homemade BSD kernel with clustering/failover features, the combination of which allowed us to operate far below competitors' costs. Acquired by another ISV in 1999, ultimately rolled into a regional BoC.
I’ve worked at some of the top computer science research labs, including Bell Labs/AT&T Research, with most of my research centering on operating systems, data storage, and networking.
In addition, I’ve had the good fortune to help many scientists in other fields apply computational techniques to their work.
Starting in the mid-80s, I have contributed to several of the most high-impact open source projects in the world. You almost certainly used some of them to read this web page.