[sage-devel] Re: Proof of P!=NP

2010-08-11 Thread John Cremona
That's funny -- I just spent 5 days on holiday with my old friend Neil Immerman, who won the Godel Prize in 1995 for proving NL = co-NL,. I don't think that he mentioned the P=NP thing once. And then I get back online and find that sage-devel is buzzing with it! John On Aug 11, 3:10 am, Bill Ha

[sage-devel] Re: Proof of P!=NP

2010-08-10 Thread Bill Hart
On 10 Aug, 19:12, Nils Bruin wrote: > On Aug 10, 10:29 am, Bill Hart wrote: > > > Just in case this is being missed here, a problem is by definition in > > NP only if it has been shown equivalent to one of the other NP- > > complete problems. > > I assume that with "equivalent" you mean "polyno

[sage-devel] Re: Proof of P!=NP

2010-08-10 Thread Nils Bruin
On Aug 10, 10:29 am, Bill Hart wrote: > Just in case this is being missed here, a problem is by definition in > NP only if it has been shown equivalent to one of the other NP- > complete problems. I assume that with "equivalent" you mean "polynomially equivalent". With your definition, P subset N

[sage-devel] Re: Proof of P!=NP

2010-08-10 Thread Bill Hart
Just in case this is being missed here, a problem is by definition in NP only if it has been shown equivalent to one of the other NP- complete problems. Thus a deterministic polynomial time algorithm for one of them implies one for all of them. It's difficult to see how you could prove NP=P withou

[sage-devel] Re: Proof of P!=NP

2010-08-10 Thread Ben Edwards
Nathann, Indeed n^9 would do us no good. My guess is if there ever was a proof of P=NP it would include a way of constructing a polynomial time algorithm for one of the many NP-complete problems. Since most of the well known NP-complete problems have been reduced to one another, it would simpl

Re: [sage-devel] Re: Proof of P!=NP

2010-08-10 Thread Nathann Cohen
> Is it actually true that if NP-complete problems could be solved in > deterministic polynomial time that Sage would be faster? No. Because knowing they are equal does not mean we can write faster algorithms. I would be amazing if the proof of P=NP contained a way to optimize at once all the algo

[sage-devel] Re: Proof of P!=NP

2010-08-10 Thread Bill Hart
Is it actually true that if NP-complete problems could be solved in deterministic polynomial time that Sage would be faster? Bill. On 10 Aug, 16:23, Ben Edwards wrote: > How much faster all of sage could be! > > On Aug 9, 9:45 pm, Nathann Cohen wrote: > > > > > -1 > > > I am hoping the two of t

[sage-devel] Re: Proof of P!=NP

2010-08-10 Thread Ben Edwards
How much faster all of sage could be! On Aug 9, 9:45 pm, Nathann Cohen wrote: > -1 > > I am hoping the two of them are equal :-D > > Nathann > > On Aug 10, 10:55 am, Ben Edwards wrote: > > > > > Just because there are a number of mathematicians on the list there is > > a potential proof of P!=NP

[sage-devel] Re: Proof of P!=NP

2010-08-09 Thread Nathann Cohen
-1 I am hoping the two of them are equal :-D Nathann On Aug 10, 10:55 am, Ben Edwards wrote: > Just because there are a number of mathematicians on the list there is > a potential proof of P!=NP out there that is being taken fairly > seriously. > > http://rjlipton.wordpress.com/2010/08/08/a-pro