This question relates to Materials science, molecular biology,
structural engineering, molecular chemistry, mapping, architecture,
medicine, transportation.  Consider a universal database is compiled. 
This database contains All Architectural data, materials data, map
data, electrical specification for wires and atomic masses of elements
and all mineral compositions and everything else we have collected as
data and - you get
the picture.  This database will put current databases to shame.  It
will be housed in a SuperComputer.  Now that all the data is in a
single space we will use filters on this data to create new things to
mankind.  For instance, we will filter out the strongest form of
structural architecture (computer prototype forms) with
superconductive materials to create the strongest electrical wire
ever.

We can call this Technique or Thesis: The Culmination of Science, The
Science of Combination.

The question is (if we assume this database is possible) what will the
filters we will use on this database to generate useful things to
humanity look like?  And to clarify:  What kinds of new things will
this supercomputer constantly be computing from our total universal
database?

To get the ultimate answer you need to ask the database the ultimate
question. A database is only a collection of data not answers. We
already have a pretty good database of worldly knowledge i.e. the
interent and libraries. It is only those that question that data that
ultimately take our understanding further forward.

true but it seems our database is fragmented. i.e. people know lots
about one career or little about lots...I think a computer combination
program constantly (24hrs a day) mingling data together in different
combinations would be a great success.  Get computers truly working
for us in a new way...?  I like your comment.

Your idea reminds me of the origin of life. When more and more
lifeless blocks get together, one day the combination as a whole
suddenly has its life. If we put all kinds of different knowledges in
one database and drive them to think, maybe this super computer will
achieve real intelligence some day. Do you mean that?

http://www.mchenrysoftware.com/SECCC%20Presentation.pdf

The algorithms of crash are are good sample of what is possible
through computer data crunching - It is not artificial intelligence
but pre programmed macros and filters - thus the intelligence still
comes from a human.

With computers we figured out how to narrow down the delta v or change
in velocity of an accident.  Delta v is like a dose of poison the
higher the change in velocity or dose of poison the more likely death
will result.  So finding delta v becomes important to designing autos
and fuselages...

The point here is we can do the same thing with all data in
combinations to find useful results...like a new material that is
stronger lighter and has colling properties of a certain criterieon so
we may create a coating that is ultra strong and lightweight and heat
deflecting for a new space shuttle "skin."
or perhaps we can find a new medicine that has the structural
properties of the bridge at firth and molecular properties of Bismuth
and Antimony but the conductive properties of niobium and copper and
also has the liquid properties of heavy water and we might use this
with a syringe to inject an epileptic patient with a strong safe
electrical conductive fluid that cure epilepsy...or alzheimers -
especially when customized to an individuals DNA sequence.

The possibilities seem endless but I do know we want this database and
supercomputers cranking out the numbers.???