> User agents must implement the SQL dialect supported by Sqlite 3.6.19.
as the only definition on what's allowed SQL, which is a no-go for a spec. The authors could have provided a specification of a required SQL dialect that's sufficient to implement it from scratch/on top of another database, but decided to abandon the effort instead.
> This document was on the W3C Recommendation track but specification work has stopped. The specification reached an impasse: all interested implementors have used the same SQL backend (Sqlite), but we need multiple independent implementations to proceed along a standardisation path.
It seems like they didn't just need a specified SQL dialect, but multiple actual implementations, but I might be reading into the wording incorrectly.