analyze

when gm's analyze games using a soft ware, like Rybka, how do they use it?

That issue about chess software or computers is very intriguing.

I imagine that those programs are made, or at least consulted, by gm´s, so my answer to your questions is "a gm consulting a gm".

But now. think about this.

a) Any chess program gives you a "best" answer for that specific positions of the pieces at that specific moment,

b) Any chess program is "good" or "better" depending on the time given the computer to analyze ("think") the situation.

c) The chess program does not know whether you prefer Bishops or Knights,

d) The program does not know if you are the type that resigns upon loosing the Queen.

e) The program considers the situation using arbitrary values given (by a human mind) the pieces according to their position and relationship to other pieces

and so on. . .

So, I think that the quality of a program depends on the reasoning of whoever made the program. . .

(I have been told of a program that always tries to get rid of all major pieces to come to a pawns ending. . .)

So I guess that consulting a program is sort of a play. . . only that that program, even being "poor", is better than an average player, because it has the advantage of "thinking", during a given very short lapse of time, 100.000 or more times than the human player. . .

. . .and, in that case, probability favors the program.

"The chess program does not know whether you prefer Bishops or Knights"

Actually, if you have certain engines, you can change those variables.  If you want a knight to be worth 3.5 points, you can set it up that way.  (Although, I doubt a GM would actually DO that.)

"Any chess program gives you a "best" answer for that specific positions of the pieces at that specific moment,"

And can also give you a list of all the moves, if you so choose.  It would actually be silly for a GM to NOT look at moves that the engine evaluates as being within a tenth of a point of the "best" move.  The engine output is used as a guide, not as an absolute "answer" as to what the right move is, unless there is one clearly best move.

"The program does not know if you are the type that resigns upon loosing the Queen."

I think most GM's would (assuming they're actually down a queen - if they have a rook and bishop for it, that's another story.)  But does it matter when analyzing the game, anyway?  If the evaluation is -9, you'll want to avoid that move, whether that's a queen or 3 minor pieces.

Very good DavidMartz. . .that is a darn good note you posted.

I don't know for sure, but I would think that the stronger engines would change the "value" of pieces slightly.  We certainly do when we play.  We've all seen good bishops and bad bishops.  We've all seen a bishop with no movement, and a knight that can reak havok.  A good engine would almost have to do something similar.