3 ms·
as far as i see a "universal relation" system is Prolog and the likes.
by trhway 2y ago
as far as i see a "universal relation" system is Prolog and the likes.
- bottom999mottob 2y agoProlog's resolution engine is more about logical inference - it will happily backtrack through different paths to find solutions. What Graham proved specifically was about relational databases: if you try to maintain consistent joins across all possible paths (aka "join consistency"), it becomes an NP-complete problem. The "Universal Relation Problem" requires ALL possible join paths to give consistent results. Prolog doesn't have verify consistency across an exponential number of join path, so I fail to see how it is a "universal relation" system.