| PLEX86 | ||
25th Anniversary of the Personal Computer" 4199Not easier...Address space estimates have to include a count of what you need to keep track of. That always gives your first back-of-the-envelop calculation. Anything with an AI would have to keep track of a LOT. Just watch yourself think. 25th Anniversary of the Personal Computer" 4202 On Tue, 15 Aug 2006 09:41:28 +0200 Perhaps - OTOH I found MP-M systems with a 6MHz Z80... And they weren't AI; they were expert systems and could not emulate an expert. (Expert was the first step of AI.) Flow charting the decision making process of an expert drove sane men mad ;-). Women had no troubles with this. I don't know if I can describe this in English ASCII. It's one of the things I do best. But that's the point; you don't know what you need in the future. Most people kept these bits hanging around in memory. Then swapping was invented; segments; files; directory trees; all of this stuff was to "store" information that you might need some day. AI was supposed to be able to make the decisions of what pieces of this bit soup was needed in advance of knowing it was needed, retrieve it, and have it ready by the time the computing process required the bits. Now, just this paragraph would show that 10x estimate is foolish.
I don't know LISP either. I never watching people while they were thinking in LISP. But I know its a very "foreign" thinking pattern. It's one that I've never studied and I always put it in a fourth category after OS thinking, compiler thinking, database thinking, and then "other" which had LISP thinking in it.
emoticon turns head to the right and gazes into the junk room that is partially painted, all torn apart, with future Unix system hidden under tarps made of towels...Slloooooooowwwwwlly. 25th Anniversary of the Personal Computer" 4200 Philip Homburg Peter Flbutt The 286 segmented memory management has been much maligned. I may be the only programmer in the world who likes it. It does have some limitations which seem odd today, but... emoticon jumps up and down in frustrated excitement I AM!!!!!!!! We were talking about all this and starting to seriously study it a year or two later. They should have done some OS programming, especially of the TOPS-10 flavors. Any AI would have to have an interrupt system driven by "events". Although how those events are determined would have been fun to do. Think about it. Any branch in a program could be an "event". That was around the time that Kalman (last name began with R but I cannot remember how to spell it--sorry, Kalman) went off. I cannot remember what that company's name was. Wasn't their machine design the one you're talking about?
Yes. I don't remember Kalman nor RDH saying anything about 10x. Everything they would talk about implied a lot more. BAH
|
||||
25th Anniversary of the Personal Computer" 4200 Alt Folklore Computers from Newsgroups The #1 Usenet Provider on the Internet
|
||||