SQL was first created or defined by IBM in the 1970's. There have been enhancements through the years, but the middleware (it is called middleware because it represents the link between the database and the user) hasn't evolved to meet the realities of the modern internet paradigm.
We are storing data, meta-data, graphics, music, and all sorts of digital stuff like we never have before. And what do we do with it? We plunk it into a database and for the most part, there it sits.
Companies have come to realize that static stored data can be monetized and contribute to the bottom line. So they purchase all of these add-on data mining and data warehouse software packages to slice and dice their data. And what do they use? SQL that hasn't changed very much from when it was conceived.
What we need now, is a serious SQL upgrade. We need functions like SELECT REGRESSION(ColumnA, ColumnB) that will help us analyze the data. We need stuff like SELECT WEIGHTED_MOVING_AVERAGE(ColumnA, 100 values). We to know if a table insert is an outrider or a fat tail. We need to know R-SQUARED of column values. We need to be able to read the contents of videos in BLOB columns. None of this has been done.
We still labor on with primitive functions in the middleware with rich client or heavy back end stored procedures to compensate. It is time that someone looked at making intelligent middleware. We need to change it from Structured Query Language to Superb Questioning Language.