http://www.dbazine.com/burleson27.shtml



Maybe not, but for once I think he's not too far
off the mark. Reckon there is more than a grain of
prediction in all that...

MOF, 2020 might be a tad optimistic: I think
2015 is more realistic.

Let's brace ourselves for the PUT - Petabyte
Undo Tablespace (sorry,folks: just cooked dinner
and had to relax...) and absolutely MASSIVE
data redundancy as a normal way of life.

I'm currently helping with an app that will reconstruct
events piece by piece in a complex workflow by logging
them, how, when and by whom, while allowing the users
to reconstruct any bit of info anywhere at any stage.
Giving them the option of taking a branch at any point
of the timeline and making that the main thread.
Multi-threads may be acceptable in the future. Need I
mention declarative RI is nowhere near enough to handle
all this? Bulk collect and squeeze the most out of
Connor's old book is the order of the day. Mark my
words: learn pipelined PL/SQL, it is mucho "kewl"!

Whatever this will all be used for, is in the realm of
the cognoscenti and the deed of non-disclosure.
Relax. Nothing sinister, quite the opposite in fact.

Welcome to the brave new world of unlimited hardware
resources and limited human capacity to use them.
We need a new paradigm: "database" is too short a
word for the engine(s) to handle all this.

See you, whenever I surface again. (who said life was
boring?)

Donald, you have a lot to answer for with this one...
Bastard!