home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cis.ohio-state.edu!zaphod.mps.ohio-state.edu!saimiri.primate.wisc.edu!usenet.coe.montana.edu!ogicse!das-news.harvard.edu!cantaloupe.srv.cs.cmu.edu!GS35.SP.CS.CMU.EDU!sgall
- From: sgall+@CS.CMU.EDU (Jiri Sgall)
- Newsgroups: comp.theory
- Subject: Re: Silence on P=NP
- Message-ID: <By6xov.Muv.2@cs.cmu.edu>
- Date: 23 Nov 92 22:41:16 GMT
- Article-I.D.: cs.By6xov.Muv.2
- References: <1992Nov21.222755.18863@sophia.smith.edu> <1992Nov23.091136.7123@fwi.uva.nl> <By6nDC.6Gp@dcs.ed.ac.uk>
- Sender: news@cs.cmu.edu (Usenet News System)
- Organization: Carnegie Mellon University
- Lines: 30
- Nntp-Posting-Host: gs35.sp.cs.cmu.edu
-
- In article <1992Nov21.222755.18863@sophia.smith.edu> you write:
- >The silence on the Swart & Gismondi proof of P=NP is deafening.
- >Will no one venture an opinion? My knowledge of linear programming
- >is insufficient to the task.
-
- To say the least, it's very poorly written. All the informal introductions
- are just a random informal statements about P and NP that could be produced by
- any undergradute student who just passed an algorithms class. In a paper on
- such results I would expect at least some nontrivial insight about the methods
- used and why should they work. Main results are stated as corollaries, no formal
- definitions, etc.
-
- The technical parts introduce precisely some linear programming problem. Then
- the argument goes along the lines "this does not work but if you modify it in
- the following way, it will". Certainly not a way in which you write a proof in
- mathematics. Also things like talking about "a point in the convex hull of a
- larger convex set" do not give an impression that the authors have a good
- grasp on the subject.
-
- My opinion is that nobody can take this writeup seriously. The papers should be
- rewritten completely so that they resemble a formal proof and than someone can
- check it. Probably the usual way would be to submit it in a journal or a
- conference like FOCS/STOC.
-
- Of course, if the proof is incorrect (which I believe is the case) then the
- best thing to do is to "publish" it as a poorly written unrefereed technical
- report - it certainly maximizes the time until the error is discovered.
- (I think the ultimate version of this technique is due Fermat :-)
-
- -- Jiri Sgall
-