At the recent ApacheCon North America 2019 in Las Vegas, the opening keynote session celebrated the 20th anniversary of the Apache Software Foundation (ASF), with key themes being: the history of the ASF, a strong commitment to community and collaboration, and efforts to increase contributions from the public. The session also featured a talk by astrophysicist David Brin on the potential dangers of AI.
The session began with a Founders' Panel, featuring several founding members of the ASF, including Mark Cox, Roy Fielding, Jim Jagielski, Brian Behlendorf, Cliff Skolnick, and Lars Eilebrecht. The panel members recounted how they became involved in the project and expounded on the importance of community to the success of Apache. Many of the founders were present at a seminal gathering which led to the establishment of ASF, when a subset of the developers in the Apache Group met in San Francisco in June, 1998. The members of the group had been collaborating via email for years before meeting face-to-face.
The founders all called out the importance of community, collaboration, and contribution. Fielding, now famous for initially describing the REST architecture, noted that a defining characteristic of the group was that it featured members from all over the world, from both commercial and academic settings, and a key to the group's success was that they were able to organically self-organize and collaborate. Skolnick noted that he and other contributors were starting businesses and needed web servers; by collaborating, each could "give a little bit, and get a lot back." Eilebrecht recalls that he was initially surprised at his selection to join the group, given that he had not made many contributions, but he related how Ben Laurie (another founder, not present) told him that "if the Apache group measured contributions based on the number of lines of code contributed, it would be poorer for it." Jagielski commented that:
Open Source and Apache have done more for me than I can ever give back, so what I try to do is do everything I can to give back to the community.
After the Panel, science-fiction author and astrophysicist David Brin took the stage to speak on the potential dangers of AI, and a possible solution. His talk, titled "Algorithm Soup," compared existing "free-floating" algorithms and programs to the "primordial soup" of 4 billion years ago that led to the initial creation of life on Earth; Brin posits that AI "life" could emerge from this soup, rather than from deliberate efforts to program AI. Brin points out that we are used to living in a "diamond-shaped" society, with a broad middle-class and smaller upper- and lower-classes, and suggests that the threat posed by AI-entities is a return to the "pyramid-shaped" society, with an autocratic AI-entity at the top and humans at the bottom. Brin proposes that a solution to this threat is to build "cell walls;" that is, separate identities for AI-entities that can be held accountable. In essence, the competition between distinct AI entities will protect humans.
Following Brin was Christopher Ferris, CTO Open Technology for IBM, who noted an increasing trend for open-source providers to change the license on their technology to make it difficult to use the code in an open source way. There is also a pattern where cloud providers will sell a managed version of an OSS package, without contributing back to the project. Ferris shared a quote from Steve O'Grady of Redmonk:
Commercial open source providers have increasingly turned to models that blur the lines between open source and proprietary software in an attempt to access the strengths of both, with the higher probability outcome of ending up with their weaknesses instead.
Ferris claimed that "technology leaders do more than just consume open source; they also give back," and highlighted IBM's many open source releases of internal projects, contrasting that with the perception that IBM is about propriety software and mainframes. He noted that IBM was central not only to helping establish ASF, but also to the Linux Foundation and the Eclipse Foundation. And while IBM has in the past emphasized the specification of standards, their new stance is that, although standards are important, open source implementation provides a common code-base that drives interoperability and helps eliminate vendor lock-in. The key to this is open governance, so that no single vendor can change the software's license arbitrarily; this reduces the risk of adoption, making the software more successful. He pointed to the success of ASF and the Linux Foundation in providing a "safe space" for developers to collaborate, regardless of their corporate interests.
Jonathan Ellis, CTO and co-founder of DataStax, spoke next about DataStax's "three pillar commitment" to the Apache Cassandra and Tinkerpop projects: code, content, and community. Introducing the code pillar, he mentioned there is a tension between public cloud vendors and companies that contributes to open source software, specifically calling out AWS, which sells Managed Streaming for Kafka (MSK) service, but does not contribute "a great deal" to the Kafka project. MSK competes with a Kafka service offered by Confluent, which does contribute to the project. Many open source vendors have reacted by changing their license agreement on some of their distributions, but there is no clear solution to the problem.
DataStax primarily contributes to the Cassandra codebase, including several features of the upcoming 4.0 release. DataStax will also soon open source a set of unified Cassandra drivers that replace their proprietary drivers, with new features such as enhanced load-balancing and Spring Boot support. DataStax is also the largest contributor to the Tinkerpop graph database and is working on an upcoming release. Ellis announced a new cloud-based product called DataStax Insights, a monitoring tool for Cassandra clusters, and will be providing a free tier to all Cassandra users.
Sam Ruby, currently president of the ASF, concluded the session with the "State of the Feather." He began by polling the audience for contributors to Apache projects, asking how long ago they made their first. When he got to contributors of five years or less, he mentioned his hope was that soon the ASF board would contain some of these "new" members. Ruby himself made his first code commit in 1999, and by 2003 he was on the Apache board of directors.
Ruby showed graphs of ASF's growth in projects and contributors, noting the steady, nearly linear, growth over the last 20 years; currently ASF hosts 334 projects with over 7,000 committers. Ruby credits the Foundation's success to its "bottom-up, community-centric Open Source process." He then concluded by presenting a summary of the ASF's activities of the past year, including fundraising, community development, and new licenses.