[Magictour] Re: [Magictour]

JC:

>It's called 'pandiagonal'.

yes.
We could also ask for the sums of the 8 cells
which a nightrider can reach in one direction on a torus.
e.g. e4-g5-a6-c7-e8-g1-a2-c3
Can this be "magic" ?

>>>I'm now quite strong with distributed projects, so I can handle all the
>>>automatic tools, except for updating the index page :-)
>
>>can't you make the participants to do it by themselves ?
>
>No, since by experience, participants tend to start the computation
>before mailing, and also because maintaining ranges takes some time.
>In this case, there won't be a lot of tours, so we may manually assign
>them, but for example, what happens if a participant with an assigned
>tour doesn't send results after 2 weeks ?

someone else might observe this and pick the same tour.
Once uncomputed tours become rare, there will be some competition about
these.

>Do you have to mail him to ask him if the computation advances,
>or do you have to wait ?
>Mailing takes time, and waiting for answers too.

I would like to have this all done by the participants themselves.

>Another way is to assign every tour TWICE, that is to 2 different
>people, so that the computation is done at least once. This is not
>a very good approach, since the 2 people could cancel their computation,
>and it's also not very motivating.

hmm, you could run the same project twice and keeping silent about the
other project to participants of the one project.

>>yes, maybe we should make it public so everybody can join ?
>
>If you wish, you can publish a link, but I doubt that there
>will be tons of public interested in this kind of mailing list.

maybe some ounces then. ATM we just have _one_ participant

>>did Fei Lu join ?
>
>No, the only subscribed are: Gunter (twice with magictour and stertenaol),
>me (twice), George, Hugues and Tim.
>
>>JC will hopefully ban the spammers
>
>I just setup the tool so that only members can post messages.
>
>>oops, OK. I see. So not open to everybody. We're still having secrets
>>at this point ;-)
>
>No, you can share the email address if you wish.
>As I said above, we won't drain a lot of people in this kind of mailing
list.
>Also, it's for the moment a temporary list to display the progress
>of the search.
>
>>this is an interesting field to fight the battle of human ingenuity
>as demonstrated on Jelliss' pages vs. brute force computer power
>as in JC's program !
>
>>Well, the computers _must_ win one day, since they are much faster.
>>But implementing complicated and intelligent algorithms to perform
>>the tasks could be more difficult than to prove or search directly
>>by human power !
>
>Yes, a typical example is the 8x12 tour. Searching that with a computer
>can require some time !
>
>>On the other hand current computer calculations seem pretty useless
>>on a historic scale as long as computers are getting ever faster.
>>Just waiting 2 years halfs the effort ! And what's two years compared
>>to the timescale of magic knight tour's history.
>
>>Probably a cooperation human+computer is most promising :
>>Design some partial-tour and the check by computer whether
>>it can be completed.
>
>My intention if we attack the 12x12 is to build a list of all 12
>or so first moves, and distribute them via a PHP interface.

what's PHP ?

>Everything can be done automatically, but of course, I have to write
>some code to make it run.
>How will that work ?
>Simply, the program will connect to http://magictour.free.fr/assign.php
>with some parameters, get back an HTML page and will extract the
>computations to do. When a computation is done, the same page is
>called with parameters to register the tour, and get a new one.
>It's very easy to maintain, but no code is ready yet (I have to
>write a PHP program, plus the client server to connect to the page,
etc...).
>2 years ago, I thought this task was impossible to do, and now, I find
>it quite easy to code :-)

is someone else doing this already ?

>>Also, it may be interesting that you ask for an update of the mathematical
>>Bible:
>>http://mathworld.wolfram.com/MagicTour.html
>>
>>... and make an entry in Sloane's EIS :
>>0,0,0,0,0,0,0,140,0,0,0, (numer of SMKTs on an n*n-board)
>
>>what about 3 dimensional-tours BTW. ?
>
>Well, let's wait for George's answer :-)
>
>BTW, since we have almost enumerated all 8x8, what is worth
>enumerating in 12x12:
>
>- magic tour
>- diagonal magic tour
>- pandiagonal magic tour
>
>Please, can you 'vote' here, so that I can start the design of the next
>project ?
>
>Thanks !

I don't feel competent to vote ATM. I need some more discussion,reading
first. With your 12*12 I'm afraid that the search-space is way
too large to get any meaningful projects.

GS



Magictour mailing list
To unsubscribe: mailto:magictour-request@ml.free.fr?subject=unsubscribe
To mail to the mailing list: mailto:magictour@ml.free.fr


Received on Thu Dec 04 14:24:18 2001