2 Comments

Very nice read !!

I recently being a huge fan of Malloy DSL and paradigm (https://www.malloydata.dev/) and more broadly how our tooling set is getting closer the human cognition rather than "machine cognition" (if this make any sense). Think of Terraform, dbt, Kestra, etc. they all bring that declarative power, abstracting "usual" transformation.

SQL has been here for decades without major improvements. There were no such need when doing OLTP, but now with OLAP it's a different beast.

I 100% agree with the fact that it's not about tools but cognition and languages. Feeling we're a good trend as dbt, SQLmesh, SDF, Malloy, PRSQL, etc. are not really new tools : just new languages (improved SQL ?)

Expand full comment

Thanks for this - love the abstraction of “what you want to achieve” from “how the compiler says you need to do it.”

This is a translation problem, where you want to be both a fluent speaker (some things best thought in SQL rather than described first) and other things require interpolation (how do I best express the idea in this language without losing the flexibility of explaining it in non language terms.)

Semantic layer ftw

Expand full comment