Dr. Dobb's Journal January 2003
Hundreds of thousands of service robots will be installed in homes over the next few years, while the market for toy and entertainment robots will exceed a million units, according to a report of the United Nations Economic Commission for Europe (http://www.unece.org/). The report cites lawn-mowing and vacuum-cleaning robots as examples of currently available technology, and predicts that window-washing robots will be the next to hit the market. Meanwhile, Britain's Engineering and Physical Sciences Research Council has funded a project to build an ironing robot, although the working prototype isn't due until 2006.
An ISO working group has released a draft of the Fortran 2000 Standard for review and commentary (ftp://ftp.nag.co.uk/sc22wg5/N1451-N1500/N1495.pdf). Fortran 2000 is a major revision, interoperable with C and with support for object-oriented programming (such as polymorphism and type extension and inheritance). Also included are enhancements for derived types, new procedures for data manipulation, floating-point exception handling, and procedure pointers. The review period will extend until December 27, 2002.
The MIT Touch Lab (http://touchlab.mit.edu/) and researchers at University College London have demonstrated the cooperative use of haptic technology across thousands of miles and the Atlantic ocean. Two scientists—one in California and the other in London—each gripped a stylus attached to a force-feedback device, and together managed to lift a virtual block in a shared 2D environment. While press reports referred to the feat as a "handshake," the effect was closer to that of two fingers touching.
The "PHANTOM" force-feedback device has been commercialized since its invention at MIT by SensAble Technologies (http://www.sensable.com/). It is possible to "poke" someone with a PHANTOM hard enough to raise a bruise. But because there is still about a 150 millisecond delay in the transmission of touch signals, the devices must be used slowly and carefully, and it will be many years before they are used in critical applications such as remote-controlled surgical procedures.
A closely watched chess match between World Champion Vladimir Kramnik and the computer program Deep Fritz ended in a draw, with two wins to each side and four tied games. The "Brains in Bahrain" challenge revisited the man-versus-machine showdown first enacted by Garry Kasparov and IBM's Deep Blue, but several distinctions were made. Kasparov pointed out after the Deep Blue match that the programmers had access to all his past matches, but he had been denied the opportunity to study the computer's game history; that the team behind Deep Blue had reprogrammed the computer between games, forcing Kasparov to continually relearn his opponent's style; and that no breaks had been allowed during the games, giving the untiring machine an advantage. In contrast, during the Bahrain challenge, Kramnik was given a copy of the program to study before the match, and the programming team was not allowed to change Deep Fritz except to refresh its repertoire of opening moves. Adjournments during games were also permitted.
Deep Fritz is not an IBM project; it was originally written in 1991 by Frans Morsch of Holland and Mathias Feist of Germany, and is backed by the company ChessBase (http://www.chessbase.com/). Now in Version 7.0, Deep Fritz is the top chess playing program in the world, according to the Swedish Chess Computer Association rankings. Deep Blue is not ranked, as IBM retired the project after the Kasparov match, but chess analysts believe the two programs play at an equivalent level. Deep Blue's specialized hardware was capable of analyzing a crushing 200 million positions a second, whereas Deep Fritz can "only" calculate 6 million moves a second, but Fritz has superior pattern-recognition algorithms and a perfect knowledge of all five-piece endgames. An analysis of the games is at http://www.brainsinbahrain.com/.
The results of the fifth annual contest of the International Conference on Functional Programming (http://icfpcontest.cse.ogi.edu/) have been announced. This year, Objective Caml has been named "the programming tool of choice for discriminating hackers." Competing teams were given 72 hours to write a program designed to compete in a multiplayer game with the objective of delivering packages to a correct destination. The winning team—Yutaka Oiwa, Eijiro Sumii, and Tatsurou Sekiguchi—implemented their robot player in 1500 lines of code.