They're also trying to make the information catalog compatible-ish as well, so that existing web-style frameworks that parse the information schema to determine datatypes and whatnot work out of the box.