Totally agree. It should have been called “Scalability and availability (even degraded) first, with as much durability and consistency as possible”. After all, that was the motivation behind the “NoSQL” movement, when databases like Oracle just couldn’t scale, even at infinite cost, so things like Amazon went down for Christmas 2006 (?) and Amazon decided to write DynamoDB. A different set of requirements created a different product. And then MongoDB came along, initially to power DoubleClick, after relational had failed there too. Neither one had anything remotely to do with SQL but with the lack of scalability in the backend relational databases - a problem that still plagues single-primary-writer relational databases to this day.

And then other non-relational databases came along, like Couch and Cassandra and Scylla - all focused (again) on things other than “SQL vs something else”.

And now they are all adding transactions and secondary indexes and all that stuff - but starting from a much more solid base of a distributed architecture rather than the antiquated monolithic architectures of the relational leaders. Those same relational leaders (Oracle, SQL Server, MySQL, and even the current leader on the dance card, PostgreSQL) are all multi-million line monoliths which are incredibly hard to distribute, scale, and make available and operate at scale - much less easy to develop and improve.

The name is just so sad