Ever since the paper on the Vietnam of Computer Science came out, I basically grew to wonder why SQL was the only language. Good? debatable. Composable? no. spreadsheet friendly, but not json friendly.
There are what, 1000 programming languages, but only one SQL? Why?
> I basically grew to wonder why SQL was the only language
You were well ahead of the pack simply by realizing that other options are possible. It is only within the last year or two that the HN crowd has finally come to accept that SQL and relational calculus are not the same thing. If you want a good laugh, check out the older posts on HN where the comments are adamant that there cannot be anything other than SQL for relational querying because it is some kind of fundamental mathematical truth.
> There are what, 1000 programming languages, but only one SQL? Why?
There are the family of languages that get lumped together under the ORM umbrella. They are quite popular – maybe even more popular than SQL. There are plenty of other languages that are less popular, but they are often impractical to use in any real-world scenario which maintains their unpopular status.
> but not json friendly.
Not standard SQL, but if we're counting extensions duality views are JSON friendly. The Oracle tax may be a tough pill to swallow, though.
I see the same thing happening in the hardware language space. There exists a large amount of gatekeeping by people who believe that Verilog/VHDL can ever only be the one true language. So research into alternative languages die. And interfaces are obfuscated behind the language.
There are other query languages out there, many of which fix the glaring issues with SQL. For example, Microsoft's Kusto Query Language (KQL) has the "FROM" clause at the beginning to enable tab-complete, you can chain multiple "WHERE" clauses without needing to use "AND", etc...
There are what, 1000 programming languages, but only one SQL? Why?