[FOM] Boiling down proofs (ignoring "undergraduates")

[joeshipman@aol.com]

I am replying late to this "closed thread" because my membership in the 
FOM list was temporariliy disabled due to oversensitivity to "bounced" 
messages from AOL.

Aitken's point about the diversity of undergaduate curricula is 
well-taken, so to clarify, I am more interested in the general 
streamlining and simplification of mathematical proofs of particular 
results, and "teachable to undergraduates" is just a shorthand for "a 
proof that is not very hard and does not require a great deal of 
prerequisite material".

Chow's examples of "theory of special functions" and Lie Theory from 
the late 19th century are pretty good; they differ from my example of 
Dirichlet's theorem in that those results have been significantly 
streamlined in 100+ years, while Dirichlet's theorem remains almost as 
hard as ever.

I have seen a short elementary proof of the Kronecker-Weber theorem 
(another Chow example) in a Galois Theory textbook by Lisl Gaal. 
However, I now understand that a later edition of that book omits the 
proof, so maybe it was erroneous, in which case that theorem (any 
number field with finite Abelian Galois group over Q is contained in a 
field generated by a root of unity) is another good example.

The 19th-algebraic geometry results Chow refers to were not rigorously 
proved until the 20th century so I don't count them. My subject is 
"boiling down" valid proofs.

The best example I know of a REALLY long proof which has not been 
simplified in any significant way is the Feit-Thompson Odd Order 
Theorem from 1960. Can anyone suggest an earlier example?

-- JS

________________________________________________________________________
Email and AIM finally together. You've gotta check out free AOL Mail! - 
http://mail.aol.com