I actually did more than this, it’s just a bit messier. This Bluesky server that I write you from was deployed the same way! I’ll probably create a public repo with that as well in the next few days!

Ahhh I see. I mean, it’s the same with their Spark package. It has some special proprietary sauce to make it faster and add a small number of features. It doesn’t seem like total vendor lock-in though, more like an added benefit for paying a vendor for a service.

Ah, I get you.

How are you considering Unity Catalog on Databricks a form of vendor lock-in? It's an open source catalog (UC) built to work in conjunction with an open source data format (Delta Lake), which supports Iceberg natively.

Me too!