Proposals

Jason Marshall jason@george.localnet
Fri, 30 Jun 2000 13:17:29 -0700 (PDT)


> 
> Jason Marshall wrote: "Now do you understand the 
> sabre-rattling?"
> 
> Easy, Jason, put it back safely in its sheath.  I 
> don't mind reading what I already know in the off 
> chance that it might provide some further insight.

I was referring more to your earlier complaints about
a poor welcome, than to my present state of mind.

> If you had rather pour through an example with 50 
> or 100 million lines of code, instead of lesser 
> number which illustrates a point, you should have 
> so stated.<g>
> 
> As to the traveling salesman problem, yes, I am 
> aware of it.  In fact at one time in my IBM career 
> I was engaged in marketing a solution that IBM (in 
> France) had developed which had the characteristics 
> you designated (hill-climbing, etc.).  

Alright.  So we've established that you have the 
requisite foundational knowledge for you to have a
meaningful conversation with most of the group 
members, which is good.  Discussions are no fun if
you have to send someone off for a month to read a
pile of books :).

But we're coming to opposite conclusions based on the
same data, which means one (or both) of us are not
grasping the situation.

So what do we have as data?
1) Exhaustive searches (searching all possible outcomes) lead to 
	optimal answers, occasionally in suboptimal time.
2) There exists a large block of problems that have prohibitively
	(astronomically) expensive search times for optimal solutions.
3) There are heuristics that can get you near optimal answers in
	optimal time.
4) The only alternative to heuristic answers is optimal answers.
5) Your processor is frequently asked to perform tasks of type #2.
6) Modern optimizers use heuristics for all problems of type #2,
	and usually use algoritms for problems of type #1.
7) You seem to be saying that 1-5 do not preclude changing #
8) Several of 'us' have claimed that #2 has been proven to prevent #7
	from being true.

#8 is going to be a little hard to disprove, since there exist
mathematical proofs to back it up.  You cannot find optimal solutions
to some problems in trivially finite time and space, and these problems
feature prominently in the task sets for Turing machines.  So either
we're mistaken about what we think you said, or you're mistaken about
what you claim.

-Jason