March 2017 Summaries

Cockroach Labs is launching CockroachDB User Groups in New York City and San Francisco on April 4th and 5th, 2017, respectively, to engage database and infrastructure experts. These inaugural events will feature Spencer Kimball, the Co-Founder and CEO, discussing the technology and inspiration behind CockroachDB, a distributed SQL database. Attendees will have the opportunity to explore its architecture and capabilities and participate in open forums to discuss questions and ideas with CockroachDB engineers. The initiative aims to foster a community that will benefit from tutorials, pair programming sessions, and access to the latest features of CockroachDB as the user groups evolve.

CockroachDB's release of the beta-20170223 version introduced a new arbitrary-precision decimal package for Go, called apd, which replaces the previous DECIMAL type implementation in CockroachDB and is available for public use. This package addresses the need for exact number representation within an arbitrarily large range, a requirement not adequately met by standard integer and float types or existing Go decimal packages. Apd offers features like panic-free operation, support for standard functions, accurate and configurable precision, good performance, and the ability to handle conditions such as overflow and underflow. It consists of two main types, Decimal and Context, which allow for a wide range of precise and efficient decimal computations. Implementing apd in CockroachDB has enhanced its performance and accuracy, achieving closer parity with Postgres, and it is expected to be a valuable resource for the Go community, with future plans to expand its functionality. The package is open-source under the Apache license and welcomes contributions and feedback from the community.

The article reflects on the importance of research in the design process, emphasizing that effective design requires a deep understanding of the world to identify problems worth solving. Drawing from personal experiences, the author recounts two attempts at conducting a study to understand developers' experiences with database solutions. The first study, which lacked focus and produced scattered data, prompted a reevaluation of the research approach. In contrast, the second study was more successful due to strategic planning, stakeholder involvement, and a more targeted participant pool. Key adjustments, such as focusing on non-decision-making developers and optimizing the interview process, resulted in clearer, actionable insights. The author concludes that research is essential for creating meaningful and desirable products, aligning with Jon Freach's perspective on design.