The New Sudoku Players' Forum

[causbrite]

Would it be possible to have the puzzles generated by the program to be given a reference number?

Something along the lines of Freecell that come with Windows. In that program, if you come across a puzzle that you found very interesting you can select that puzzle again by typing in the number and starting.

Have a look at Freecell if you don't know what I am on about. You can select from game 1 thru to game 1,000,000.

I think it would be a worhtwhile edition to the program and on a positive note, maybe have the program give a code to those who solve every puzzle available. Challenge anyone?

[Animator]

Personally I have mixed feelings about it...

It's good since you can go back later, but ofcourse you can always save the puzzle and go that one (something you can't do with freecell)

But on the other I don't think that it would be a good idea... simply because it is a big limiation...

As in, try to calculate how many possible games of freecell exists (I think it is 52!, which is 8 * 10 ^ 67)... now compare that number with 1 million... (not to mention that earlier version of freecell had a limit of 32000)

Would it really be good to limit the number of possibilites that dramticly?

[RFB]

By coincidence I was thinking about this earlier today and think I have come up with a scheme to generate a unique character stream to concisely identify a puzzle.

I will explain later, of course it all depends on Wayne liking the idea

[Guest]

I'm not entirely sure what this discussion is about but I was wondering if it was similar to a problem I seem to be having.

Is it possible for the program to print out the filename of the puzzle it creates and I save? That is, when I dub a puzzle into Excel I've set it up to print out the filename (&[File], in the header) I've given the puzzle so that I can identify the thing and print it out later (in the event that I go wrong). Given that it takes me days sometimes to do a Fiendish puzzle I don't always remember what filename I originally gave it (especially when I've scrambled my brains trying to solve it!) so that I can reprint it. Writing it down - or rather remembering to write it down - directly after printing and before solving, is tedious and I don't always remember to do it.

Is that what you're talking about? Should I post this somewhere else? Any ideas at all?

Luna

[lunababy_moonchild]

This one was also mine, having forgotten to login.

Luna

[RFB]

A puzzle reference number would be helpful for:

• quick entry of printed puzzle from paper or book into program
• allowing sharing of a program generated puzzle
• Allow discussion of puzzles without needing to type entire grid
• correlate puzzles from papers around world

To uniquely identify a puzzle we need to identify each of the clue numbers.

The most obvious scheme is to list the row/col/value for each clue which would lead to a long number string - 60 digits for a puzzle with 20 clues which is impractible.

It would be possible to encode each clue into a single printable character but this would still result in a 20 character string that includes upper and lower case letters, numbers and puctuation characters so would be hard for a user to type.

So I am coming to the conclusion that a practical refernce number is not possible.

[Peter Morris]

I've been working on this myself. I'd very much like to see it implemented but doubt it will work.

It's a fairly common computing algorithm (2-way encoding) but not always easy to implement (sliding puzzles - easy; chess - hard; bridge - impractical). I've been able to reduce the 81 board to a forty character string using numbers and single case alphas or 30 characters using a broader character set. However, it's not generic and varies quite a bit depending on the initial number count and size of grid. For a standard 9x9, I think we're looking at a complex key of length 30 so that's not likely to be a solution.

Of course, initial pattern symmetry may help reduce this but, although Pappocom uses symmetry, not all puzzles have to be symmetrical so we have to exclude that.

Still thinking but with increasingly less hope...

[lunababy_moonchild]

Why does it need to be that hard? The books manage to number the puzzles, seemingly easily, and while it's fair to say that the books have a finite number of puzzles why can a simple numbering system for the Pappocom generated puzzles not be incorporated?

For example, I've seen on this Forum people numbering puzzles by the numbers in the first row. Since each (Pappocom) puzzle is meant to be unique, surely this is enough to generate a unique reference point - without number-crunching ourselves out of existence?

Luna

[LarryLACa]

See related Wishlist Topic: Index of Published Sudoku's. There may be others.

When restricted to published (PappoCom) or generated puzzles (by a given program), this should not be a difficult feature to add.

As it provides a distintive value-added feature to leverage Pappocom's position, I'm surprised it's not already there.