3 ms·
Relational algebra is a much closer foundation for SQL than ZFC (or plain ZF, because for finite sets the difference doesn't matter). You can say SQL has ZFC as
by kccqzy 8y ago
Relational algebra is a much closer foundation for SQL than ZFC (or plain ZF, because for finite sets the difference doesn't matter). You can say SQL has ZFC as its foundation but that's not very useful. In the same way you can say ZFC is the foundation of everything in mathematics. SQL doesn't even use anything beyond naïve set theory because we are dealing with finite sets, let alone the full power of ZF.
- neokantian 8y ago> In the same way you can say ZFC is the foundation of everything in mathematics. Yes, technically, that is indeed true of every Turing-complete axiomatic theory. You could conceivably even use the SKI combinator calculus as the foundation of everything in mathematics. Still, something tells me that a database query language in widespread use, based on the SKI combinator calculus is rather unlikely to emerge (certainly not at this point). Hence, since ZFC is one of the many alternative, equivalent foundations of everything in mathematics, picking ZFC still amounts to making a choice, no matter how small ... > SQL doesn't even use anything beyond naïve set theory ... Agreed. However, we also have to consider what situation we came from. David Childs objected to the use of custom, application-specific code, data types and functions, because these custom structures could (more easily) be represented by one, general abstraction. Given the fact that back then, people were not even using naive set theory, his considerations were clearly a step forward in the right direction.