So it’s not cheating if you scan the puzzle into a computer, click “Solve”, and fill in the squares simply by copying them down off the screen? That seems like a pretty toothless definition of cheating.
I think I answered that in my very next sentence (fragment) - because those are the rules that make it more fun for me. If someone else got the most enjoyment by clicking “solve,” that’s what he should do.
We’re throwing around words like “rule”, “official”, and “cheating” here as if they were strictly defined for an informal activity like solving a puzzle.
Official Sudoku tournaments actually exist, and there is an on-line league that has actually published some rules. Surprisingly, their rules do not allow use of paper and pencil, probably because they fill everything out on-line. Whetehr or not you consider this activity and the published rules to be silly, they at least offer a clear process to determine if something is official or cheating.
Regarding the more informal activity, it might be best to expand the question: For what reason does someone attempt to solve a Sudoku puzzle? It should be clear that the methods used to solve a Sudoku puzzle should at least support this reason if the solver wants to solve the puzzle legitimately. If, for example, someone claims to do it because they want to test their own mental reasoning skills, but ends up feeding the puzzles into a computer or just waits for the answer in the next day’s paper and copies it into the grid, the solver is probably cheating. On the other hand, if the solver’s stated motive is simply to see what the solution is (perhaps just to understand how puzzles of this type work), these methods are legitimate. Yes, frustration can often get the better of us, which may lead us to, say, ask for a hint from someone who has already solved the puzzle. But we clearly do this because our motive changes, so again–at least with respect to our initial motive–getting a hint is a kind of cheating.
In general, people solve puzzles to test their mental and logical abilities with respect to the puzzle. Memory is not what the Sudoku is trying to test, so use of paper and pencil to aid memory is not cheating.
The rule about pencil and paper seems OK, but it looks like it’s also against their rules to say “Fidel Castro”!
To me, it is clear that the database understands nothing about chess. It couldn’t explain how to play a particular position, or even look ahead more than one move.
It just retrieves the best move in a particular position.
The number of legal positions is estimated at 2*10^43 (Fox + James: Complete Chess Addict).
This is a large number :eek: , but it’s not infinite.
Once you can get all these positions into computer memory, chess is on its last legs.
A superb post - well done that man! 
The only point I would make is that chess programs won’t try to analyse what humans find difficult. Firstly human players vary considerably in their styles, and secondly it would be difficult to program.
Sadly brute force and no ‘intelligence’ computers will triumph. 
What if it was a computerized Sudoku program that you yourself wrote? What if it was a computerized Sudoku program that you yourself wrote that didn’t just brute-force it, rather, it came up with the best set of human-understandable logical steps it could come up with?
Another way to think about it is that the word “solve” only has meaning if it communicates that meaning to others. Thus, you only “solved” it if what you did would be accepted as solving by the Sudoku community as a whole.
It’s not cheating. It’s not fun. But it’s not cheating.
As I said above, as a game designer the most useful definition of play that I’ve found is Huizinga’s. In his theory play requires three elements: a bounded play space, a lack of real-world consequence, and a rule set that constrains action.
The rules of play in any game only exist by the consensus of the players. If two chess players decide in advance that pawns move two spaces instead of one they’re not cheating. They’re playing a variant. Cheating is when one player breaks the rules that he agreed to. Since sudoku is played solitare each person can define his own rule set. When I play sudoku I think it’s fun to limit myself to no extra marks on the paper and no erasures. I feel like I’m cheating if I use any other techniques to solve the puzzle. Other players may define their rules differently.
However, there are limits. You could construct a game around two players throwing chess pieces at each other, but it would not be a variant of chess. You have to preserve some essential elements. Filling in the grid to meet the no duplicate number criteria is essential to sudoku. If you don’t fill in the grid at all, or fill it in with numbers that meet different criteria, you’re not playing sudoku anymore. (Although you may be having fun.)
The problem with the one button solution isn’t that it’s breaking the rules (cheating), but that it’s eliminating them entirely. Play depends upon having a rule set to push against. Eliminate all constraint and the tripod of play collapses. That’s why it’s not fun.
In this case you’re not playing sudoku. You’re playing a different “game” that you “win” by creating a program capable of solving a sudoku puzzle. The programming environment and the programming language themselves define the “game board” and “rule set”.
Well, presumably the looking-ahead is built into the database. That is, at each position, the database builders analyzed every possible move, then every possible response, then every possible re-response, out to a certain number of moves, and selected the “best move” as the one that leads to the most favorable outcomes several moves into the future.
I’d say that as long as you are not in a competition with set rules, you can solve the puzzle in any damn way you wish, may it include pencils, excel sheets or computer programs that solve it for you. Its not until you use the result for something, like winning that competition, or maybe make yourself look smarter than you actually are, that it can be seen as immoral to use those extra tools.
In the end, Sudoku is just a game, for you to enjoy as you see fit, cheating doesnt really apply in every day life, IMO…
My answer is based upon tradition rather than philosophy.
I did the puzzles decades ago in Dell puzzle magazines before they were called sudoku. For the more difficult ones, I used a pencil and wrote tiny notes. To me, that wasn’t cheating, but just a memory aid. To me, cheating would have been peaking in the back of the book just for a second to get a few more numbers in the grid. I had no computer, but when it came to other math-related puzzles, I considered using a calculator to be cheating.
The point of puzzles for me is working out the logic, rather than testing my memory, so I would take notes outside the supplied boxes when it came to logic problems too. If the human isn’t supplying the logic part, what’s the point of playing at all?
As Will Shortz said of crosswords, “it’s your puzzle.” Solve it however you want. There’s your non-philosophical answer. If you’re enjoying yourself, continue using the software.
The philosophical puzzle of whether or not you are solving the puzzle is also yours to enjoy as you see fit. 
Not a problem…unless you extrapolate that out to mean those pencil markings, at some point, will become you.
If true, I’ll need a big eraser and a correction marker…
(Kidding, of course ) 
You’ve obviously never played Full Body Tackle Chess.
I understand the last two elements, but what about the first? Is it really required? Or is it just a practical constraint that most games don’t bother avoiding?
Consider some game where the players are trying to find the largest integer they can that fulfills some set of criteria. (I don’t know of any game like that, but it’s conceivable.) There’s no limit to the set of integers. Or how about the game Nomic, where the moves of the game include changing the rules of the game. That’s a peculiar case, I’ll admit — but there it is. There are hardly any limits at all to where the game might go.
Games that are set in a physical space are obviously bounded, because people’s stadiums, yards, and table-tops are always bounded. But a more abstract game played on an infinite “board” could still be fun, and fit the definition of “play”. Or so it seems to me, anyway.
I just want to comment on Bytegeist’s reference. Pete Suber, in my opinion, is one of the two greatest logicians in the modern world, the other being his nemesis, Alvin Plantinga. Seldom have there been two such contemporary giants at opposite ends of the philosophical spectrum. It was their battles that drew me into the study of logic. Thanks, Bytegeist.