Forums

Cyclone wrote:

Any chance I'm going to see in on here?

Certainly, that's what this thread is for. There are still many other requests that we haven't implemented though (as well as some genres that have not been requested but that we have implemented), so it's hard to say anything about when you're going to see it here...

I'd love to see Tapa puzzles on your website.

sojaboon wrote:

I'd love to see Tapa puzzles on your website.

Thank you.

I made an implementation of this genre, but what was not immediately clear to me is if the numbers necessarily show all lengths of neighbouring sequences. The example puzzles I came across seemed to suggest this, but the texts did not. Could you clarify this?

If anyone wants to help to speed up the Tapa launch, just drop me an email with a bitmap of your puzzle + solution. Then we will make sure it is transferred to your account.

Yes, they show all lengths. For example, f there's a 3 in a square, then there's one section 3 connected cells coloured around the square and the rest is empty.

If there's time for a new latin square variant can we put kropki in? I prefer the normal kropki puzzles over the kropki sudokus, but that might be my taste.

Latin square rules and all instances where consecutive numbers are adjacent are marked with a white dot and if one number is half of the adjacent number are marked with a black dot.

I second Tapa. It's a nice puzzle type.

Tapa will be launched tomorrow morning.

Kropki has been added to my TODO list.

I would love to see Fillomino, especially since I'm trying to author some of those.
writing an interactive form of the puzzle may be a bit difficult since borders, as well as grid cells need to be edited to solve this puzzle type.

@rantash: I think we can make it work one way or another. If you want, you can already work on them in the studio by using the Neighbours genre. This might speed up the launch. Just leave a note under your puzzle stating it is supposed to be Fillimino, then we will convert them to the Fillimino format once we've decided upon the exact implementation.

If I were you, I might make the Fillomino applet where the solver can input numbers or boundary lines, but only the boundary lines need to be correct to count as solved. I've played with other Fillomino applets, and personally, I find that without the ability to add boundary lines myself, it's a lot harder to solve. I personally always solve these on paper by adding both numbers and boundaries. *shrugs*

It's interesting that mathgrant write to check only for the boundries for correctness, I though to check only the numbers; however I often solve fillominos on paper and forget to fill some numbers. I guess boundries are better, but only if they would also be automatically deduced from neigbours with different numbers; they don't have to be filled in visually, they may only be inferred for the checking procedure.
Another suggestion for the applet would be that you can drug a number into other cells continously to fill other cells just like you would color cells in the LITS applet.

I think that Akari would be cool.
Here are the details from Nikoli.

This is my spin-off for "puzzle sudoku".
The solution for a puzzle sudoku can be used for "reverse engineering".

3324422 <- How many thick lines in the column

2134657 1 <- How many thick lines in the row
4725361 2
1567423 4
7253146 2
3612574 3
6471235 4
5346712 1

The problem?
Form the (seven in this case) pieces by drawing the thick lines.
Each piece must have all numbers (1-7), of course.
I believe this is harder (more guessing) than solving the original puzzle sudoku (not published yet)...

Solution:
1111122
3331222
4431225
4433355
4466557
6465577
6667777

Jfo wrote:

This is my spin-off for "puzzle sudoku".
The solution for a puzzle sudoku can be used for "reverse engineering".

Cool. We currently have an unreleased genre where you dissect a grid (not necessarily a square or even a rectangle) into polyominoes containing all different symbols exactly once (we use letters at the moment). Adding the possibility to give hints regarding the number of line segments in a row or column sounds like it could work.
Do you happen to have some of these ready?

"Reverse engineering" when puzzle #2329 is the solution would be:

22433222

4256713 1
7145236 5
2673145 2
1462357 3
6531472 4
5317624 4
3724561 2

Now, you *only* have to draw the lines between the pieces to solve it, just like Seethrough. And yes, I have MANY puzzles...