I think one big issue is that professors certainly don't care to explain how a certain building-block really affects something in the real world. For e.g. as Jo

I think one big issue is that professors certainly don't care to explain how a certain building-block really affects something in the real world. For e.g. as Joel explains, how having all rows in a DB as fixed length dramatically improves performance vis-a-vis XML (because .. you know CPU operations :-) ) really will get students think through the strategy of "Oooo!Shiny! Let me use this" prevalent in lots of companies.