Announcement: Be excellent to each other.


Caravel Forum : Caravel Boards : General : Dealing with the new tar (Man it's tricky.)
1
Page 2 of 2
New Topic New Poll Post Reply
Poster Message
Dex Stewart
Level: Smiter
Rank Points: 355
Registered: 01-19-2007
IP: Logged
icon Re: Dealing with the new tar (0)  
Just to be nitpicky. This:
Mikko wrote:
 XX    
XXXXXX                 x  
XXXXXX transformed to xxxx
   XX                   x 


is cuttable.
04-10-2007 at 09:01 AM
View Profile Send Private Message to User Show all user's posts This architect's holds Quote Reply
Maurog
Level: Smitemaster
Avatar
Rank Points: 1501
Registered: 09-16-2004
IP: Logged
icon Re: Dealing with the new tar (0)  
Care to elaborate how you cut it? As I see it, you are left with a 2x2 block no matter what you do. ;)

____________________________
Slay the living! Raise the dead!
Paint the sky in crimson red!
04-10-2007 at 10:08 AM
View Profile Send Private Message to User Send Email to User Show all user's posts Quote Reply
Maurog
Level: Smitemaster
Avatar
Rank Points: 1501
Registered: 09-16-2004
IP: Logged
icon Re: Dealing with the new tar (+1)  
Speaking of smallest shapes that are cuttable in all three tarstuff forms, I think this one is the prettiest:
 XXX
XXXXX
XXXXX


____________________________
Slay the living! Raise the dead!
Paint the sky in crimson red!
04-10-2007 at 10:55 AM
View Profile Send Private Message to User Send Email to User Show all user's posts Quote Reply
Dex Stewart
Level: Smiter
Rank Points: 355
Registered: 01-19-2007
IP: Logged
icon Re: Dealing with the new tar (0)  
Seeing how convinced you are, I looked back. Was he by any chance talking about Tar? I'm asking because I thought it was gel... :blush
04-10-2007 at 12:34 PM
View Profile Send Private Message to User Show all user's posts This architect's holds Quote Reply
Maurog
Level: Smitemaster
Avatar
Rank Points: 1501
Registered: 09-16-2004
IP: Logged
icon Re: Dealing with the new tar (0)  
We're talking about Gel. Here are the only scenarios I see for that shape:
 XX                  ..                           __                __
XXXXXX   Cut NW ->  .,XXXX   After NW cut SW ->  __..XX  or SE ->  __XX..
XXXXXX              ..XXXX                       __.,XX            __XX,.
   XX                  XX                           ..                ..

                     ..                           __
         Cut NE ->  XX,XXX   After NE cut SE ->  XX_... 
                    XX.XXX                       XX_.,.
                       XX                           ..
Cutting the initial shape SW or SE is mirror reflection of the above. In all cases you are left with a 2x2 block.

____________________________
Slay the living! Raise the dead!
Paint the sky in crimson red!
04-10-2007 at 01:12 PM
View Profile Send Private Message to User Send Email to User Show all user's posts Quote Reply
Briareos
Level: Smitemaster
Avatar
Rank Points: 3516
Registered: 08-07-2005
IP: Logged
icon Re: Dealing with the new tar (0)  
Maurog wrote:
We're talking about Gel. Here are the only scenarios I see for that shape:
I guess this is simply semantics - yes, of course said shape is cuttable, but it's not destroyable, i.e. you can't cut it until there's nothing left...

____________________________
"I'm not anti-anything, I'm anti-everything, it fits better." - Sole
R.I.P. Robert Feldhoff (1962-2009) :(
04-10-2007 at 01:21 PM
View Profile Send Private Message to User Send Email to User Visit Homepage Show all user's posts Quote Reply
Yellow_Mage
Level: Master Delver
Rank Points: 267
Registered: 05-19-2004
IP: Logged

File: Fun With Gel.hold (1.9 KB)
Downloaded 55 times.
License: Public Domain
icon Re: Dealing with the new tar (+1)  
I've figured it out (at least partially - probable good reason why it's forbidden), but I'm not very good at explaining things.

When you cut, exposed surface pieces will nearly always become babies. In clearing Gel, this isn't particularly important. It's better to focus on the new surface that is made; do they become edges or flat pieces (or inner corners). This is why you can cut Gel one way but not another.

I had something going with that you could only clear Gel by only cutting inner corners on adjcent diagonal tiles, but there are exceptions.

So I made a room with Gel so people and poke about and ruminate on this because my BRIAN muscle isn't working very well at the moment.

(plus is much easier messing about in the editor than trying to draw in text)

____________________________
"Sit and daydream, and watch the changing color of the waves that break upon the idle seashore of the mind." - Henry Wadsworth Longfellow


Click here to view the secret text

04-11-2007 at 01:01 AM
View Profile Send Private Message to User Send Email to User Show all user's posts Quote Reply
bomber50
Level: Smitemaster
Rank Points: 872
Registered: 09-18-2006
IP: Logged
icon Re: Dealing with the new tar (+1)  
Eveyone here is missing a very important thing about gel!

Click here to view the secret text




[Last edited by bomber50 at 04-12-2007 12:01 AM]
04-11-2007 at 11:59 PM
View Profile Send Private Message to User Show all user's posts This architect's holds Quote Reply
silver
Level: Smitemaster
Rank Points: 915
Registered: 01-18-2005
IP: Logged
icon Re: Dealing with the new tar (0)  
bomber50 wrote:
However, your sword must be different from your mimics...

To do that, you need disarm tokens. Place a mimic. Then use the disarm token.

great puzzle mechanic that could be used in a variety of places :)


____________________________
:yinyang
04-12-2007 at 12:02 AM
View Profile Send Private Message to User Show all user's posts This architect's holds Quote Reply
KevG
Level: Smiter
Avatar
Rank Points: 333
Registered: 08-16-2004
IP: Logged
icon Re: Dealing with the new tar (0)  
The disarm token isn't strictly necessary. You could achieve the same affect with a clone potion or a patch of oremite.
04-12-2007 at 09:45 AM
View Profile Send Private Message to User Send Email to User Show all user's posts Quote Reply
Maurog
Level: Smitemaster
Avatar
Rank Points: 1501
Registered: 09-16-2004
IP: Logged
icon Re: Dealing with the new tar (-1)  
Keep it simple, people. The mimic was already in the room when you got there.

____________________________
Slay the living! Raise the dead!
Paint the sky in crimson red!
04-12-2007 at 09:46 AM
View Profile Send Private Message to User Send Email to User Show all user's posts Quote Reply
Daneel
Level: Delver
Rank Points: 35
Registered: 04-12-2006
IP: Logged
icon Re: Dealing with the new tar (+2)  
Another question about gel...
We seem to have a reasonable understanding of how to analyze destroying gel, we can ask is there an algorithm to determine whether all gel can be destroyed (ignoring mimics and death). Unlike the other tarstuff, I have a feeling that this problem might be NP-Complete.

I have been thinking about reducing circuit SAT to gel destroyability. I haven't gotten any gates working yet, but I do have a wire. Consider 4 rows/columns of gel (with some corner to cut far away). In order to clear it, you need to cut either all of the black squares or all of the white squares, but since you cannot remove ANY two adjacent squares (after removing the first, the second becomes uncutable), and since you need to cut one of each pair of adjacent squares as shown:

XXXX
XSSX
XXXX
XXXX

in order to clear the square above it, you must pick either light or dark squares to cut, hence a wire.

I think you can split wires, but am not entirely sure (this would have to create new corners) to prove NP-Completeness, all we need is a branch, a NOT, and an AND/OR/NAND/NOR gate. Any ideas?
04-19-2007 at 08:15 AM
View Profile Show all user's posts Quote Reply
Daneel
Level: Delver
Rank Points: 35
Registered: 04-12-2006
IP: Logged
icon Re: Dealing with the new tar (0)  
I think that I can prove that gel destroyability is NP-Complete.

I already explained my wire. To branch, have a wire with another branching off on one side. The far edge ensures that the original wire is of one type. Additionally make sure that there are only entrances of one type on the branch (you can only cut into it at dark squares or only at white squares). This means that although the branch is not necessarily the same as the main, there are only errors in one, specified direction.

To get a gate, take a large box with corners of a particular color (say white) that must be destroyed. Attach a number of wires to it in such a say that you only provide opportunities to cut black squares. Therefore, to clear the gel, you will need one of the incoming wires to be white. This produces a gate of the form

A OR B OR C OR ...

or of the form

(NOT A) OR (NOT B) OR (NOT C) OR ...

Lastly, we have branches with error, and need to branch wires without error. So we can branch off a wire A, a wire which caries A/1 (either A or 1) or A/0. What we can do is the following:

Have two wires, A,B. Have branches A/1, A/0, B/1, B/0. Connect these branches with gates

(NOT A/1) OR (NOT B/1)
A/0 OR B/0

The first ensures that either A or B is 1. The second ensures that either A or B is 0. Hence A = NOT B. Repeating this twice gives us another copy of A.

Hence we manage to reduce planar circuit SAT to gel destroyability proving NP-Completeness.
04-19-2007 at 11:58 PM
View Profile Show all user's posts Quote Reply
schep
Level: Smitemaster
Avatar
Rank Points: 865
Registered: 03-01-2005
IP: Logged
icon Re: Dealing with the new tar (0)  
Daneel wrote:
I think that I can prove that gel destroyability is NP-Complete.
How do you attach an "input" to a wire end? I can't imagine a way the player could choose either color to start cutting, and still manage to clear everything nearby.

EDIT: No, wait, this is easy.
. X X X X
X X X X X ...
X X X X X
. X X X X 


Also, to make a proper proof out of this, you'd need to demonstrate turning and crossing these "wires".

Come to think of it, are you really talking about the problem "Can this shape of gel (no obstructions) be cleared by one creature and sword?"

EDIT 2: Your "gates" sound more like final predicates than something you can fairly use in later expressions. A gate in this setup would need two input wires and an output wire.

[Last edited by schep at 04-20-2007 01:14 AM]
04-20-2007 at 12:52 AM
View Profile Send Private Message to User Send Email to User Show all user's posts This architect's holds Quote Reply
Daneel
Level: Delver
Rank Points: 35
Registered: 04-12-2006
IP: Logged
icon Re: Dealing with the new tar (+1)  
OK. So crossing and turning wires we do not do directly.

For turning wires, we note that our procedure for duplicating wires can produce duplicates at different angles from the originals.

Crossing wires is hard. We do this on the predicate level. Essentially we are reducing to the problem PLANAR circuit SAT.

OK. So we have NOT gates. We can produce an AND gate (i.e. C = A AND B) using the following predicates:

A OR (NOT C)
B OR (NOT C)
C OR (NOT A) OR (NOT B)

(and noting that these can be arranged in a nice planar manner)

To simulate crossing wires is hard (note though that we only simulate crossing wires)

The idea is to produce a device with inputs A,B,C,D with A and B on the left and right, and C and D coming from up and down that is satisfiable if and only if A = B and C = D.

We will have four intermediate arbitrarily chosen values W,X,Y,Z that are arranged vertically between A and B.

We construct something that looks like the following:

.
..........C...............
..........|...............
...X AND (A OR (NOT B))...
|---------|--------------|
A.........X..............B
|---------|--------------|
|..X AND (A OR (NOT B))..|
|.........|..............|
A.........C..............B 
.

We compute the X AND (A OR (NOT B)) as (X AND A) OR (X AND NOT B), so that it can be done in a planar manner. We additionally attach similar gadgets with the expressions

Y AND ((NOT A) OR B)
NOT [ Z AND (A OR (NOT B))]
NOT [ W AND ((NOT A) OR B)]

Note that all of these expressions are equal to both C and D, so C=D. Note that if A = B, the (A OR (NOT B))'s and ((NOT A) OR B)'s are always true, so W,X,Y,Z can be picked to make this work out. On the other hand, if A is not equal to B, then one of these expressions will fail. For example

C = X AND (A OR (NOT B))

is satisfyable unless A = 0, B = 1, C = 1. The other expressions rule out the other three combinations of (A,B,C) where A is not B.

Hence this gadget allows us to simulate wire crossing.

[Last edited by Daneel at 04-20-2007 03:52 AM]
04-20-2007 at 03:50 AM
View Profile Show all user's posts Quote Reply
schep
Level: Smitemaster
Avatar
Rank Points: 865
Registered: 03-01-2005
IP: Logged
icon Re: Dealing with the new tar (0)  
Well, now we're almost there.

We still need to prove that any satisfiable logical expression maps to a bunch of gel which actually can be cleared. I played around in the editor and proved that all three theoretically possible 'states' of the following wire junction are clearable:
. . X X X X . .
. X X X X X . .
X X X X X X . .
X X X X X X . .
X X X X X X . .
X X X X X X . .
. . X X X X . .
. . X X X X . .
This is Daneel's "branch with error" (in one color; to get the other, put the little bump in the other corner), or put another way, an IF-THEN (A->B) predicate.

However, I haven't been able to come up with the simple OR predicate, the other basic piece we need (in both a "straight" and "right angle" variety). Everything I've tried seems unclearable or clearable in only one way, not two.

If anyone is still reading but not following all the logic stuff, the challenge in that last part is: Find a shape of gel with two 4-wide "wires" attached, such that
a. The shape has a corner like
. . .
. X X
. X X
on a given checkerboarded-floor color, say light.
b. The shape has no cuttable corners on light squares.
c. It's possible to clear the entire shape by cutting both wires on light squares, AND possible cutting wire A on dark squares and wire B on light, AND possible cutting wire A on light squares and wire B on dark.

Another thing that's important to note in the proof: Assuming we find appropriate OR predicates, we don't need to worry about the restrictions of parity on the placement of the shapes, because we can just place the OR predicates where we want and let them determine the positions of the junctions like I drew above, which can be anywhere and still favor whichever "color" we want, by moving that one square of gel.

[Last edited by schep at 04-21-2007 12:32 AM]
04-21-2007 at 12:23 AM
View Profile Send Private Message to User Send Email to User Show all user's posts This architect's holds Quote Reply
Daneel
Level: Delver
Rank Points: 35
Registered: 04-12-2006
IP: Logged
icon Re: Dealing with the new tar (0)  
OK, my solution to (A OR B OR C OR D) is as follows:

.
. . . . . . X X X X . . . . . .
. . . . . . X X X X . . . . . .
. . . . . X X X X X . . . . . .
. . . . X X X X X X X X X . . .
. . . X X X X X X X X * X . . . 
. . X X X X X X X X Y X X . . .
X X X X X X X X X Y X X X X X X
X X X X X X X X Y X X X X X X X
X X X X X X X Y X X X X X X X X
X X X X X X Y X X X X X X X X X
. . . X X Y X X X X X X X X . .
. . . X * X X X X X X X X . . .
. . . X X X X X X X X X . . . .
. . . . . . X X X X X . . . . .
. . . . . . X X X X . . . . . .
. . . . . . X X X X . . . . . .
.

You need a positive input from one of the wires in order to clear the *'s. I have figured out a design (though not tested it) for how to clear the whole region given some positive input. You reduce the number of cases you need to look at to 3 by noting the symmetry and requiring that you always cut along the Y's.

If you want to ignore some of the inputs, you can set them to false (by starting the wire like this:

EDIT: (the original was wrong and could not be destroyed)
.
. . . . .
. . X X X
. . X X X
. X X X X
. X X X X
. . . . .
.
), or by just not using that input in the diagram.

[Last edited by Daneel at 04-21-2007 05:39 PM]
04-21-2007 at 01:30 AM
View Profile Show all user's posts Quote Reply
schep
Level: Smitemaster
Avatar
Rank Points: 865
Registered: 03-01-2005
IP: Logged
icon Re: Dealing with the new tar (+1)  
No, a constant-valued wire ending should look like one you've already started cutting:
. . . . .
. . X X X
. . X X X
. X X X X
. X X X X
. . . . .

Maybe I'll see if I can get your great big OR working later.
04-21-2007 at 01:24 PM
View Profile Send Private Message to User Send Email to User Show all user's posts This architect's holds Quote Reply
Neathro
Level: Delver
Avatar
Rank Points: 41
Registered: 04-24-2007
IP: Logged
icon Re: Dealing with the new tar (0)  
I think i'm still asleep...
is this DROD or ST2??? about wires, fuses???

____________________________
This is not a signature, it is a
collection of randomly selected
words. Teaspoon.
04-27-2007 at 09:51 AM
View Profile Send Private Message to User Send Email to User Show all user's posts Quote Reply
NiroZ
Level: Smitemaster
Rank Points: 1302
Registered: 02-12-2006
IP: Logged
icon Re: Dealing with the new tar (0)  
Neathro wrote:
I think i'm still asleep...
is this DROD or ST2??? about wires, fuses???
DROD. It helps if you read from the start.
04-27-2007 at 12:05 PM
View Profile Send Private Message to User Send Email to User Show all user's posts This architect's holds Quote Reply
1
Page 2 of 2
New Topic New Poll Post Reply
Caravel Forum : Caravel Boards : General : Dealing with the new tar (Man it's tricky.)
Surf To:


Forum Rules:
Can I post a new topic? No
Can I reply? No
Can I read? Yes
HTML Enabled? No
UBBC Enabled? Yes
Words Filter Enable? No

Contact Us | CaravelGames.com

Powered by: tForum tForumHacks Edition b0.98.9
Originally created by Toan Huynh (Copyright © 2000)
Enhanced by the tForumHacks team and the Caravel team.