2023-12-26

A few weeks ago was my undergraduate advisor Norman Ramsey's retirement party. Norman taught two main classes: Machine Structure and Assembly Language Programming and Programming Languages, which were both the most demanding and rewarding courses in the Tufts CS curriculum. Many students, myself included, found these courses formative, and the right of passage was the bedrock for the culture of the department and a number of enduring relationships.

Fellow students, TAs, advisees traveled from afar to share anecdotes of Norman's teaching. Many of us, myself included, attribute our success in our endeavors since school to Norman. This man was a legend, if I haven't made it clear. There were tears, jokes about compilers, and stories of overcoming imposter syndrome.

Someone at the event mentioned that they were designing a domain-specific query language in Rust for their work. This sparked an idea for me - the query languages that I've encountered to-date have been designed for table or graph-like structures of data. As my mind becomes more immersed in the domain of meteorology, n-dimensional data structures have stood out to me as underserved by the design of languages and file formats. After the party, walking through campus, I wondered to Alden (who also traveled to the retirement party), what if weather data had a programming language or query language expressly designed for its challenges? What would it look like? For now, I'm calling it wxql.