The New Sudoku Players' Forum

[dyitto]

How would a text format of this puzzle look like?

Here's my suggestion on this, it allows to specify land, water, and also the undefined cells (which could become land but maybe not).

Code: Select all

2;+;3;.;.;.
+;-;.;.;.;.
-;.;4;.;.;.
.;.;.;5;.;.
.;.;.;.;.;.
.;.;.;5;.;7

Legenda:
. (dot)    undefined
- (hyphen) water
+ (plus)   land
n (number) numbered cell, also belongs to land

The first puzzle of this topic is taken as example.

[simon_blow_snow]

How would a text format of this puzzle look like?

Here's my suggestion on this, it allows to specify land, water, and also the undefined cells (which could become land but maybe not).

If you read my walkthroughs, you will see that I already devised a version of text format for these puzzles using '@' for land and '#' for water.

But if '+' & '-' make more sense for you then I can adapt, it just takes a few clicks to perform a "replace all". But I don't use ';' to separate cells, since for me they in a way affect my sense of perception to the horizontal connectivity between cells.

I just used '@' & '#' because they are more unusual, so can be replaced easier in a long text file of walkthroughs. If I try to do a "replace all" on '+' & '-', I will probably mess up the text in the walkthroughs.

[simon_blow_snow]

In this link you can find the puzzles made by Naoki Inaba, who came up with the genre.

http://inabapuzzle.com/honkaku/island.html

Hey Para, have you tried this game from Naoki's site:

http://inabapuzzle.com/anime/chika.html

Turn all the squares to yellow within the limited number of steps. Hover your mouse to move the bot. Click the bot to change the colour of the square it stays on as well as the adjacent squares in front, on the left and on the right, in respect to the direction of the bot's eyes. (So a maximum of 4 cells will change colour when the bot is clicked.)

(Hovering the mouse within the game area doesn't count as a "step", but clicking any mouse button, pressing any key or hovering in and out of the game area all do.)

I think stage 5 is unbeatable using pure logic. I had to use a small computer trick to crack it.

If you don't want to go through the first 4 stages (which are all kindergarten level easy) yourself, here is the password to go straight to stage 5:

Password for stage 5: Show

505050

[dyitto]

If you read my walkthroughs,

I should have done that

you will see that I already devised a version of text format for these puzzles using '@' for land and '#' for water.

But if '+' & '-' make more sense for you then I can adapt, it just takes a few clicks to perform a "replace all". But I don't use ';' to separate cells, since for me they in a way affect my sense of perception to the horizontal connectivity between cells.

I just used '@' & '#' because they are more unusual, so can be replaced easier in a long text file of walkthroughs. If I try to do a "replace all" on '+' & '-', I will probably mess up the text in the walkthroughs.

OK what does your text format look like if numbers > 9 occur?

[simon_blow_snow]

OK what does your text format look like if numbers > 9 occur?

Normally I will follow the hexadecimal convention, e.g. A=10, B=11, ... and then it can be extended to G=16, H=17, etc.


Currently working on stage 4 of Naoki Inaba's Island puzzles. Stage 3 was an absolute beast. I have a feeling it is even tougher than stage 4.

It is also the first time I see numbered cells stick together. From Para's puzzles I thought they should always stay apart.

[dyitto]

It is also the first time I see numbered cells stick together. From Para's puzzles I thought they should always stay apart.

Also zero could occur as number theoretically

[simon_blow_snow]

Also zero could occur as number theoretically

You're right! Show's how flexible the rules of this nice puzzle genre are.

BTW I've finished stage 4 of Inaba's Island puzzles. I take back my earlier statement that stage 3 is tough than stage 4. Turns out stage 4 is a bigger beast despite its smaller size. If anyone likes to discuss about it we can do it here. Just no time for any walkthrough for now.

[dyitto]

BTW I've finished stage 4 of Inaba's Island puzzles. I take back my earlier statement that stage 3 is tough than stage 4. Turns out stage 4 is a bigger beast despite its smaller size. If anyone likes to discuss about it we can do it here. Just no time for any walkthrough for now.

I can't see those puzzles unless I solve some stages

[simon_blow_snow]

I can't see those puzzles unless I solve some stages

dyitto, I've sent you a PM. Check your inbox.



Have you also tried the animation game I mentioned earlier?

http://inabapuzzle.com/anime/chika.html

Many of these games are quite nice in that they (mostly) take under an hour to complete, while giving you a nice thorough stimulating mental exercise during that short time.

[dyitto]

Here are two of the easiest Island puzzles from Para's blog, just so you can get a feel of them:



I will post the walkthroughs for these puzzles later, that is after I finish writing them!

I've written a very basic solver, and a backtracking algorithm.
For the first of these puzzles it needs 7 seconds to establish that it has one unique solution. For the second, it takes at least longer than 5 minutes and I'll try another time to have that run finished.

[simon_blow_snow]

I've written a very basic solver, and a backtracking algorithm.
For the first of these puzzles it needs 7 seconds to establish that it has one unique solution. For the second, it takes at least longer than 5 minutes and I'll try another time to have that run finished.

I wouldn't be surprised. Like I said these puzzles are highly intuitive, so perfect for human mind but very difficult to teach a program to solve them mechanically.

You might like to read my walkthroughs and see if you can teach your solver any new concrete technique to speed up the solving process.

[Para]

BTW I've finished stage 4 of Inaba's Island puzzles. I take back my earlier statement that stage 3 is tough than stage 4. Turns out stage 4 is a bigger beast despite its smaller size. If anyone likes to discuss about it we can do it here. Just no time for any walkthrough for now.

I can't see those puzzles unless I solve some stages

http://www.janko.at/Raetsel/Naoki/Airando.htm

Here's the puzzles without the stages.

[Para]

OK what does your text format look like if numbers > 9 occur?

Normally I will follow the hexadecimal convention, e.g. A=10, B=11, ... and then it can be extended to G=16, H=17, etc.


Currently working on stage 4 of Naoki Inaba's Island puzzles. Stage 3 was an absolute beast. I have a feeling it is even tougher than stage 4.

It is also the first time I see numbered cells stick together. From Para's puzzles I thought they should always stay apart.

It's not a requisite. The puzzles that I made that I like best just are with clues apart.

[dyitto]

I've written a very basic solver, and a backtracking algorithm.
For the first of these puzzles it needs 7 seconds to establish that it has one unique solution. For the second, it takes at least longer than 5 minutes and I'll try another time to have that run finished.

I wouldn't be surprised. Like I said these puzzles are highly intuitive, so perfect for human mind but very difficult to teach a program to solve them mechanically.

You might like to read my walkthroughs and see if you can teach your solver any new concrete technique to speed up the solving process.

It was speeded up by checking if two landcells can be found that are definitely separated by a water line - which was actually quite easy to program.
The second puzzle was solved in 3 hrs. Checking that the solution is unique was at least ready when I looked the next morning.

Here's my own first puzzle:

Code: Select all

.;.;.;.;.;1
.;.;.;.;.;.
5;.;.;.;.;2
.;.;.;.;7;.
1;.;.;.;.;.
.;.;.;.;2;.


[simon_blow_snow]

Here's my own first puzzle:

Code: Select all

.;.;.;.;.;1
.;.;.;.;.;.
5;.;.;.;.;2
.;.;.;.;7;.
1;.;.;.;.;.
.;.;.;.;2;.


Nice simple puzzle which captures the spirit of the rules for beginners!

Solution for dyitto island #1: Show

Code: Select all

#####1
#####@
5@@#@2
@#@@7#
1###@@
####2#

I should also try to write a short program to verify uniqueness using brute force. Should not be too hard I suppose?