Despite our industry's habit of treating mathy whiteboard things as indicative of one's "programming talent", real application of type, set, and graph theory are not tested nearly as much as they end up being in practice, and I think that developers' reticence to get deep into SQL is symptomatic of the weaknesses there.