Author: Dann Corbit
Date: 16:03:20 04/17/00
Go up one level in this thread
On April 17, 2000 at 17:58:12, Tom Kerrigan wrote: >On April 17, 2000 at 13:37:23, Dann Corbit wrote: >>On April 17, 2000 at 05:37:08, Tom Kerrigan wrote: >>>On April 17, 2000 at 04:20:57, Jeremiah Penery wrote: >>>>On April 17, 2000 at 03:06:21, Jouni Uski wrote: >>>>>I have got impression, that Deep Blue has much more complex and better >>>>>evaluation function than PC programs (it had >8000 different properties). Is >>>>>this knowledge now lost? Can Hsu publish his code? Can it be used in PCs? Or may >>>>>be DB code is suitable only for it's paraller hardware? Bob, Can You explain. >>>>> >>>>>thanks Jouni >>>> >>>>FWIW, I think they should at least do _something_ with the code/machinery of DB. >>>> It truly would be a great loss if none of it was ever made available for >>>>examination. >>> >>>I doubt DB's algorithms are any better than those of top PC programs. This has >>>been discussed before, thought, and I don't think a good conclusion can be >>>reached without knowing more about DB. >>> >>>I think the real novelty of DB is the chip design. Hsu got the rights to the >>>design from IBM. He had some plans to make commercial DB chips for PCs, but I >>>think those plans have fallen through. He probably won't make the design public, >>>though, because it might be comercially viable in the future. >> >>If the chip is patented, then the information is publicly available. And when >>the patent expires (7? 9? years) anyone could make one. However, in that time >>frame, I suspect it may be a moot point. > >I doubt it's patented. Don't you have to have some unique, new methods to get a >patent? DB is just an evolutionary step from DT. You would be amazed at what gets granted a patent. I think (for the most part) software patents are ridiculous. I'm not competent at patent law, as far as even knowing what *ought* to be patented, but I have seen plenty of examples that were clearly absurd and yet granted, which shows me that the patent office is really 'without clue' when it comes to this sort of thing. From my view, since it is illegal to patent mathematics, I think all software patents should be denied, but I'm probably much too radical on that standpoint. As far as a chip is concerned, patents make more sense to me, but only marginally so. In any case, I would not be the least bit suprised if the chip were patented. A patent is (by far) the strongest protection for an invention. Basically, you have a complete monoply until the patent expires. Here's an interesting article on Hsu's chip: http://www.eetimes.com/news/98/1022news/ibm.html And another article: http://www.edtn.com/embapps/emba028.htm
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.