Define "chess engine"

My other hobby, besides games, is programming. So now that I'm back into playing Chess (thanks to an old friend who brought me to this site), I thought I would write a Chess program. I figured I could write an analysis board type program, with some extra features. But then as I started working on the program, I realized I wasn't sure what features I could add.

I want something that you could legally use on this site for online chess. The rules here state you can't "use any chess engine that analyzes your specific position." The relevant Webster definition of analyze is "to study or determine the nature and relationship of the parts."

Now, the analysis board on Chess.com already informs you of illegal moves. If you try to move a rook diagonally, it doesn't let you. It seems that showing all the legal moves for a given piece would be OK (in fact, GameKnot's analysis board does this). But what about a feature where you could click on a square and see all the pieces that could move there? Would that be "determining the ... relationsips?" What if you could see a summary of that information for all the squares at once?

Then there's the pawn counts for taken pieces. Chess.com provides this also, using the classical 1/3/3/5/9 scale. But there are other scales out there, some of which take into account the bishop pair and whether pawns are isolated or doubled. Are those relationships such that providing those alternate pawn counts would be analyzing the position? That seems okay, but a full blown evaluation function taking into account king security and center control seems over the line. But I'm not really sure where that line is.

I was going to allow the storing of lines of analysis so you could look at them later. But what if the program could alert you when you were making a move similar to a stored line of analysis? That seems to be a database use, which is mentioned as OK in the rules, but then I got confused by the restriction against tablebases, which seem to be databases.

If anyone has any ideas, or can point me to a discussion of this somewhere I couldn't find with the search feature, I'd appreciate it.

a chess engine is simply a program which tells you in some sort of way what move is good for you to play, and affects your judgment in what to play in a position directly.

what you describe is cheating at the rate of taking the OTB game home to your own board so you could analise it without your opponent in the room, in online chess it is perfectly legal, as that is exactly what you so in the game, if you are unhappy of the chess.com analysis board it's a perfectly fine decition to build your own virtual board with features like:

legal moves.

matereal imbalances (displaying things like two minor pieces vs a rook).

displaying the opening names or even an opening database.

a link to game archives with the same position.

you can think of other things, but nothing concerning the future of the game, only the present or the past...

I've always believed that the chess.com rules were pretty vague and a little messed up.  Databases are pretty much the same thing as computers in my mind.  Both systems tell a player what the best move is, and If you are not deciding what the best move is, then it isn't really all that fair.  Trying to push the boundries shouldn't really come into play it we just made firm boundries: For correspondence chess a player should only be able to use the tools provided to every player (ie. analysis board).  I come to this site so I don't have to play computers, and playing against a database or a book kind of ruins a game.  Save those for between games. 

 Only 6 to 8 moves ?!  I know the openings I play regularly well into the middlegame and so do many " class " players. Opening books now also often go well into the middle game as well. Take a look at the Marshall counter attack in the Ruy for example and it goes beyond move 30 in most opening " Bibles" which are meant to be just opening references. A book dedicated solely to the Marshall goes even deeper......sometimes to recurring endgames . Whats interesting about data bases is that their percentages ( thus their choices of the "better moves" are based on human tournament experience, and sometimes many of those players are below master level. I sometimes run such positions through a strong engine and the engine often disagrees completely with the information given by the data base !

Your other points aside, I'd like to correct this mis-assumption. Neither databases nor computers (engines) can tell you what the best move is for certain. People tend to think "well, Fritz says this move so it's best." But that same Fritz program is losing games to Rybka all the time. And Rybka today will be losing to next year's best program. A chess engine is not a perfect player.

oh, it sure depends..
6-8 moves is really not a lot. depending on how familiar a person is with the opening, a 'C' class player might be able to go more than 15 moves deep.

besides, this site allows the use of opening books and databases. ;)

Exactly, I doubt any Class 'C' player would have trouble going 6-8 moves deep in a Ruy, KID, QGD, Sicilian, semi-slav, QGA, etc. I mean 6-8 moves is barely scratching the surface nowadays when it comes to opening theory. 

Displaying all possible moves is a lot more than the Analysis Board does. With the AB, you have to attempt any move before the system prevents you.

Showing all captures, all checks, all undefended pieces (or all defended pieces), possible pins or skewers would give a great deal of information visually; it would certainly constitute calculating a future best move.

Man do i feel stupid!!!!!

Kepler> If your opponent allows you to follow a book or database line for more than 6-8 moves on this site your opponent is either a computer or a master of some description.

Reb> Only 6 to 8 moves ?!  I know the openings I play regularly well into the middlegame and so do many " class " players.

Kepler> I am guessing that NM next to your name means you are a master of some sort. If so you just confirmed what I said.

Many of my OTB games against class A players stay in-book for 9-12 moves. Furthermore, using opening books is permitted here. I would not attribute staying in-book 8+ moves on chess.com to cheating, and imho such is not even a great indicator of skill level.