I don't think chess will be solved any time soon, and paradoxically, computers, in combination with humans, might be part of the reason.  For instance, see the Spanish-like line I've devised in this thread:

http://www.chesspub.com/cgi-bin/yabb2/YaBB.pl?num=1335889682/

A lot of it was largely done with the help of Fritz; sometimes specific moves, but sometimes also general conclusions such as "if White's queen were on e2 then after Bxc6/...dxc6/Nxe5/...Qd4 would recover the pawn due to Black's ...Nf6."

Or this thread where using Chessbase I find a not-unreasonable move in the Gruenfeld/Exchange, used to defeat Portisch in the early 60's, and adopted once by Smyslov later in the decade, and yet played only a dozen times, practically a thousand times less than the main line.

http://www.chesspub.com/cgi-bin/yabb2/YaBB.pl?num=1336013487

GM Flear is considering ...Qa5 instead of the main line this month, so why not ...Nd7.  Certainly it might warrant hundreds of attempts rather than only a dozen.