SQL is simpler to understand for majority of users. It's easier to get started without learning a tool chain, programming best practices, etc. that could present a challenge to new users. SQL is also great at representing relationships between datasets and developing business logic transformations tends to be simpler and easier to understand. Oftentimes, when using another programming language you end up with many modules, imported libraries, code hacks and optimizations. It all can very quickly make it difficult to read.
One example is dbt. They started as writing simple SQL models. With the introduction of Jinja, majority of models look nothing like SQL anymore. You need to visualize your model to understand relationships. It took the beauty of SQL and mucked it.
At Upsolver we built a streaming+batch ETL tool that lets you build data pipelines in SQL. We did it because it's easier for non-data engineers to get started, easy to version and maintain as code and easy to automate (not that you can't do this with other languages). The same goes with kSQLDB, Materialize and even Spark and Flink use SQL as a way to simplify onboarding for non-developers.
it's a great write up and the context issue seems interesting, however it's from Feb 2010 according to the last paragraph and a lot has changed in V8 since
One example is dbt. They started as writing simple SQL models. With the introduction of Jinja, majority of models look nothing like SQL anymore. You need to visualize your model to understand relationships. It took the beauty of SQL and mucked it.
At Upsolver we built a streaming+batch ETL tool that lets you build data pipelines in SQL. We did it because it's easier for non-data engineers to get started, easy to version and maintain as code and easy to automate (not that you can't do this with other languages). The same goes with kSQLDB, Materialize and even Spark and Flink use SQL as a way to simplify onboarding for non-developers.