You're right Dr. Bad. I knew upping the max value (200 above) would allow larger bubbles to be found, but the time goes up exponentially. And we're still talking single-gate bubbles. I've gone so far as 3 gates with TWX, but it requires lots and lots of time. I'm sure there are better algorithms for multi-gate bubbles, as mine just used simple elimination. I've pondered it in the past, came up with an idea or two, but it would take a month for me to test and implement them.

I think you answered the key question though; only an uninterpreted language or method would be feasible time-wise to locate large bubbles.

FC, please tell me that you stumbled upon that 1800 sector bubble. Because if you have an effective way to search for them, I'd love to hear it. 1800 sectors... incredible.

+EP+

quote:Originally posted by Father Cajone

Now Doctor Bad...when you going to poke your head out of retirement and come back and play with us a bit? Got a big draft game coming up where you could teach a lot of us some of your devious tactics!


'fraid it aint gonna happen anytime soon- been a bit too much on at work to have time for serious tw these days.

I ran a few tests with twassist in 5K bangs. The maximum bubble size I had was 326 sectors, some bangs max size was 6 sectors. Size did not affect the time it took to produce bubble lists, it was always 7 or 8 seconds. No clue what method it uses unfortunately, dont have source code.

I got to give it to Rev, His Cim reports have a secludeds report, have found a certian section of the universe seperated by 1 sector, that doesn't show in his bubble report, not in every game, but quite a few, I'm in a unlimited game that had a long truce, ( yea it sucks now, no one can invade anyone )But One corp owns one section and we own another, then there is the rest of the universe. I would post one for an example, but its too big...lol

I don't know much about Gold Bubbles, but is it true that all 2-way exits will pass through a tunnel of at least length 1? If so, you can dramatically reduce processing by first identifying all tunnels that do not dead-end. Then, for each of the 2 sectors on either end of the tunnel, perform the breadth-search that EP describes. However, you need a second criterea for stopping: all paths lead to a second tunnel. You've then identified a bubble, of some size. Larger bubbles all share a tunnel or gateway in common, so you can then construct all bubbles > 200 sectors by simplying checking for bubbles that have identical exit points. You should also be able to allow for arbitrary numbers of exits, since only tunnel sectors are a legitimate exit.

I have not actually coded this out for Gold bubbles because I'm not sure about that tunnel assumption. The savings should be 10-100 fold depending on game settings for bubbles < 200 in size (compared to EP's method of searching every sector), and finding larger ones costs almost nothing. Even if it's not exactly a tunnel requirement for bubbles, there must be some straightforward requirement to identify the only sectors that possibly could be an exit.

quote: but is it true that all 2-way exits will pass through a tunnel of at least length 1?
You lost me on that question.
quote: If so, you can dramatically reduce processing by first identifying all tunnels that do not dead-end.
By definition, a tunnel is a string of sectors (2 way warps) that has at least one Entrance and one Exit. If they do come to a D/E, then they could be classified as either;
A) Bubble
B) Dead End.

(Correct me if I am wrong) I just finished writing some tunnel scripts, and their seems to be 2 definitions of a tunnel. A natural tunnel that has 2 Entrances and 2 Exits (Excluding Gapped Tunnels that contain a sector with a Backdoor) and a One Way tunnel with 1 Entrance and 1 Exit.
BTW - Anyone interested in this Tunnel script, contact me. I would also like to further discuss the possibilities of any Abnormal Tunnel's I may have overlooked.
quote: have not actually coded this out for Gold bubbles because I'm not sure about that tunnel assumption. The savings should be 10-100 fold depending on game settings for bubbles < 200 in size (compared to EP's method of searching every sector), and finding larger ones costs almost nothing. Even if it's not exactly a tunnel requirement for bubbles, there must be some straightforward requirement to identify the only sectors that possibly could be an exit.
I have been pondering this same question since I first posted this thread, without resorting to looping through incremental sectors and using a breadth-search. Is there a way to define a "Gate"? I don't think so, assuming that the "Gate" is in fact not a tunnel, and that it can contain any number of warps. I beleive the only way is to examine each sector then follow those warps until they lead back to the original Sector in question. Although I'm still plugging away at alternate methods of searching for bubbles.

Zoso