From: thunk on 31 Mar 2010 05:23 On Mar 31, 3:24 am, Andrea Dallera <and...(a)andreadallera.com> wrote: > Hei, > > I gave a look to your github wiki: while still I have no idea what > you're aiming to do at least it looks that you're trying for real. > I am in no position to suggest you anything but personally I'll be > getting more interested in this project of yours when 1) i can get some > code that i can execute and see what it does from start to end 2) all > this talk about Ruids and Boids boils down to something concrete: I've > got some (not much) experience in data mining and pattern matching and > what i learned is that what sounds cool often does not work. > > Bye and good work! > > -- > Andrea Dallerahttp://github.com/bolthar/freightrainhttp://usingimho.wordpress.com hej, its a living breathing / processing something on my machine, what can I say, i'm not typing in imagined output - want it from 1000 patients - i'll post it up in 2 minutes - its real. my software was gathering mostly numeric data from factories around the world, really, from cmms and such since the middle 80's to the middle 90's on DOS based real-time apps i/we wrote back then. then came windoz and then i ran screaming to Smalltalk that IBM had just "re-re-introduced into VisualAge" and that seemed cool to me. Never connected with the Smalltalk community from NE Wisconsin - I think Smalltalk was rather dying even it as it was the best thing I've ever seen in some ways - imagine a IDE and complete (100%) OO after having been forced at gunpoint into buggy c++. no don't - if you can understand - it may make you cry. but we also gathered comments usually associated with statistical events. I wrote / authored all kinds of reports based on that to about 1995 and it was interesting work that was worthwhile and fun and rewarding.... and I think we were about the only folks doing it too. However, the point of this post is that "i have pissed up a few ropes" (a colorful expression i reserve for serious discussions) and i can tell you with almost complete personal confidence that this is no such thing. which is wierd because the heads on approach blew up complexity wise - and that was like I have said - to try and resolve faulty data on the fly throwing complextity head on at complexity only seemed to grow it. and that is what it is - bad data needs to get resolved sooner or later right? The "Ruids" don't worry about this because the "SwarmSender" must guarantee that this has already been done - in this "new" approach. But that's cheating?! ya. I changed domains. I now have excellent data, a tad more complex than i'd like but just excellent - it is nicely organized into sentences which I put into my "Satz" (almost just a hash but has some rules - sorry but it seemed to deserve a unique word). Cute thing about THIS is that the computer gets to play with data it can handle heads on - and we humans can see the reconstitution of the Satz (technical) into "Techlish" which is "better than english or any natural language as keeping the ands & ors and double negatives all straight. I'm CHEATING, so to say. You guys can resolve the tough stuff - like working with impossible units or missing units and implied units - i have no such thing in my little world - and I can tellya why- because I'm working with Drugs as a CLASS not as an instance - and the FDA makes sure that the instances spec out to the Class - in the computer world of CPUs and such there is no such regulation - and AMD / Intell let their marketing guys write their respective garbage and its going to take a real expert to sort that out. I'm not a real expert, I don't even know one up here - so I had to tuck my schwanz betw my legs and move to where i didn't need to be an expert. crystal clear? BUT there are things about Computer systems like continual change and I still see real value to Blasting a shopping basket with 1000's of teamed "Ruids" to figure out 1. if its anything like a "system" 2. if something obvious might be missing 3. if the stuff that logically mates will mate 4. if there isn't a better system for the money (or whatever) and if his accumulated fan noise is above / below expected, if the power supply seems adequate, and most of all if the thing should work....... and the "Ruids" can be contributed by computer geeks - that was how "the lightning stuck my head"! - i wanted to allow outside contributions and THAT has been accomplished beyond what I expected and one of the reasons i stay genuinely excited about this project. It leverages the whole thing - hook this up to a Forum like PCMech and you rather can own a Greek Island or two, and invite me over. OK? I wrote to PCMech and some other such domain champions and they are not ready to "see" this yet. and, ahh, well, geee, but it isn't the easiest thing to communicate about.... not yet anyway - we need to agree on symbols for this stuff and the RU in Ruid trips my trigger - and that seems related to this forum somehow, just vaguely. thunk
From: thunk on 31 Mar 2010 05:47 hej, and the beauty of the shopping basket app(s) go deeper... e.g; CompUSA (that I was scraping) and NewEgg provide all the details, daily price changes, open box specials, shipping and details - all that need to be done is to "get into those "shopping baskets"" - then only HTML type output. then take 5% of the money SAVED and you have a living or two. see what i'm trying to say? the urRuid app that I could not finish ==> and I don't give up easy. This thing, i've been tell'n a few people, could be hooked up to forums for input (not to much cost, actually they provide a real service) and Spree or some such eCommerce system. pure ruby goodness put to work.. a real nice loop to close. thunk
From: thunk on 31 Mar 2010 06:11 eeeiii! what are those Ru'id things do'n in my Shopping Basket!??? i had it worked out that cables were a really nifty objects with nested "plugs" and that checked if they would mate with whatever which was also a "plug" and all like that... it would have worked too, there are air flow calculations and heat dissipation factors, not to mention assembly options and lots of real value to be added imho and my favorite scenario repeated I don't know how many times, was the PC tech at a bench in Island reporting a cpu/mem glitch that manifests as something that needs to be tweaked in BIOS (or some such stuff that occupies these guys) would automatically show up to some guy in S. Africa years later with the same computer fingerprint. that seemed pretty exciting to me. One "Ruid" filled out on a handy could do that.... then there are aquariums, and kayaks, and other artifacts of male passion... (and I love kayaks too - I and I know a real world class designer in this domain - the one's that go on expeditions - i'm into "skin on frame" myself - that's how they started you know.) thunk
From: Intransition on 31 Mar 2010 08:15 Turning Test? On Mar 31, 3:25 am, thunk <gmkol...(a)gmail.com> wrote: > ok, some clarity arrived: > > for trivial games one "helperclass" each. one '?'method per potential > move. chess would require maybe 6 helper classes or so. the "ruids" > play n play players. the swarms are organized differently than i have > done but no big deal. each swarm would handle a new move by any > player. and no big deal. > > ok. > > but the rules of the game are determined by the ruids. the ruids > could replace an existing rule (not just elbow in an be accumulated). > this means that the game rules could be changed as the games are being > played. the rule changes could be trivial, or they could be > controlled by some meisterspieler that is adding interest to the whole > mental menagerie of male bemastering. > > so vying to change the rules - "politics" - could become part of the > whole RuidScape. > > now that seems pretty amusing and somewhat less trivial.
From: thunk on 31 Mar 2010 09:17
On Mar 31, 7:15 am, Intransition <transf...(a)gmail.com> wrote: > Turning Test? no Turing test. thunk ps at the current stage i'm authoring "system ruids" that i consider to be "processing" oriented". reminds me of "boot strapping" in the sense that I feel I am only laying the foundation of this domain for what is to come. very few surprises happening. srand used to seed each batch... no surprises. there is a "black box" recording every attribute acquisition asf. as said, the hook to the helper_class methods can be used for anything. i'm keeping it all as simple as I can for now. you slingers can take it to new places, I'm clint eastwood. |