Blame Fermat for favoring publicity over proof. --- Actually, there may be a much larger problem, at least in the math community. Only the really important theorems get the intense scrutiny of Hilbert's Tenth, 4CC or FLT, which often require years of revisions to finally get the complete proofs correct. I can only imagine that many less important published 'theorems' might not be able to withstand this level of scrutiny. This isn't to say that the theorems are wrong. There are often very good empirical/experimental reasons for believing that the theorems are true, so it is probable that most of the published theorems are true, even though the published proofs leave a lot to be desired. Those of us in the computer software community face this issue every single day. We work for a tirelessly nitpicking slavemaster, who is quite adept at finding that bad 'corner case' in the most inopportune/embarrassing moment. We can only hope that computer proof-checking will soon advance to the point where math journals will routinely required a certificate of proof-checkability before the paper can be accepted. At 06:01 PM 5/28/2013, Robert Baillie wrote:

here's bailey/(jon) borwein's take on this and other issues: http://www.huffingtonpost.com/david-h-bailey/scientific-research_b_3340682.h...