test files?

anyone want to post some test files... i think my algorithm works but i would like to check it against someone elses.

explain your algorithm... Flip the greatest value to the top and than flip the whole stack, is NOT the fastest. That's the reason why I quit this contest. Now I'm building a script that searches in the Google database with SOAP

You quit the contest because it was challenging?

do you want input files or the actual algorith script?



you won't get my algorithm ;o)

I think the algorithm part is WAY beyond coding now. Just making an algorithm now is a task in itself. I'm also not sure if I'll do this contest. Matt's idea of running a "dual contest" with a GUI part too is novel, but I think that detracts. I would rather there be two "normal" contests. So, possibly have another contest that's not so algorithmic. I mean, the last contest I knew how to approach it: a DFS tree.

This contest, I have no idea what to do. I think one could practically write a research paper on the topic of prefix reversals before they'll beat the guaranteed 2*n-move algorithm.

In other words, if someone does come up with a new algorithm, there's no reason to even code it. Just publish it!

Then again, it's possible someone's going to do a search tree for a smaller test-case (only 50 or so discs is still quite a large number though...) and the search result will be enough to beat everyone else who is only using the guaranteed algorithm

It's possible the move-sums will be something like:
990, 1002, 1002, 1002, and everyone else has 1002 because they're all using the 2*n-move way.

I have been able to get #flips < n using an optmized version of my original algorithm. the only problem i have is that sometimes its off by one, i verified my flips using my gui counterpart.. but i cant figure out what is the case that its off by one.. thus i was asking if someone had a complex stack that they knew 100% what the shortest # of moves is i would like to check against that to verify my code is working before i submit it and bother matt with the task of verifying that the method by which i am counting my flips is correct. about the gui part of the contest i hope matt will focus on the coding of the gui and not the appereance... that is if some person has a bloated up 4bazillion line flash movie that used a 3d cook flipping pancakes would be just as plausible to win as a 100 line pure html+javascript version, but like he said.. that part of the contest is all subjective. I also dont think there is a 1 step method to know the exact # of flips by just looking at the inital stack... so you would have to actually run through the flips and what not and make your script intellgient enough to know what is the best flip to make. maybe matt you could post a couple test files by which we can test our script.