home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.theory.cell-automata
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!wupost!uwm.edu!zazen!doug.cae.wisc.edu!umn.edu!csus.edu!netcom.com!pdh
- From: pdh@netcom.com (Phil Howard )
- Subject: specification of rulesets
- Message-ID: <=07mx3-.pdh@netcom.com>
- Date: Sun, 16 Aug 92 00:19:45 GMT
- Organization: Netcom - Online Communication Services (408 241-9760 guest)
- Lines: 94
-
- One way I could specify a ruleset for dealing with a rectangular universe
- of cells with 8 neighbors (4 orthogonal, 4 diagonal) would be to supply
- a bit string of 512 elements. The 8 neighbors plus original cell could
- form an index into the bit string to select the state of the new cell.
-
- However what I really want is something simply for specifying the simpler
- rulesets.
-
- In the case of Conway, there is no difference between the effect of
- orthogonal neighbors and diagonal neighbors. However I don't see why
- rulesets could be specified that did indeed have such differences.
- For example a rule might allow a cell to live with 4 neighbors if
- at least 2 of them are diagonals. It might have a cell born with
- 2 parents if they are orthogonal and opposite, 3 otherwise.
-
- Even more bizarre rulesets could even have a directional bias as in
- having the cell above and above-left count as two cells when alive.
-
- The more complicated the ruleset, the more complicated the specification
- is expected to be. Conway's rules should be quite simple to specify,
- as well as other rulesets that merely change the numbers.
-
- Here is one ruleset I am wanting to try out. Consider how such a rule
- might be specified. It certain CAN be in the 512 element bit string,
- but I think the bit string is too much for it.
-
- BEGIN RULESET I WANT TO TRY
-
- Cells will be born from 2 parents if the parents are in exact opposite
- positions from the new cell position, and there is only one set of
- parents attempting to create a new cell there. For instance:
-
- these will create a new cell in the center:
-
- o.. .o.
- ... ...
- ..o .o.
-
- however these will NOT because the parents are not opposite:
-
- o.. ..o
- ..o ...
- ... .o.
-
- note that in this case, no cell is born in the center for the same reason
- above, although one will be born on the left side:
-
- o..
- ..o
- o..
-
- in these cases, NO cell will be born because 2 or more parents are
- trying to birth in the same place and that is a conflict:
-
- oo. .o. ooo
- ... o.o o.o
- .oo .o. ooo
-
- however the conflicts only affect the position, not the parents, so
- in this case we will get NO cell in the center, but one on the left
- and bottom will be born:
-
- o..
- ...
- o.o
-
- A side note here, if the conflict affected the parents themselves, such
- as a parent encountering a conflict not being able to birth anywhere
- else, we could see the "speed of light" jump by up to 3 times. The
- calculations would also be more than a neighbor indexed lookup could
- handle.
-
-
- For cells already existing, they will live on if the neighbor FACTOR is
- between 2.5 and 5, with diagonals counting as 1 and orthogonals counting
- as 1.4. So that means:
-
- 0 1 2 3 4 <-- orthogonals
- 0 . . X X .
- 1 . . X . .
- 2 . X X . .
- 3 X X . . .
- 4 X . . . .
- ^
- |-diagonals
-
- Which definitely treats the 2 types of neighbors slightly differently.
-
- END OF RULESET I WANT TO TRY
- --
- /***********************************************************************\
- | Phil Howard --- KA9WGN --- pdh@netcom.com | "The problem with |
- | depending on government is that you cannot depend on it" - Tony Brown |
- \***********************************************************************/
-