#eulora Logs for 20 Aug 2016



August 20th, 2016 by Diana Coman
diana_coman: finally got building to 600! [09:49]
mircea_popescu: wd! [10:06]
mircea_popescu: dc destroyed my nice bot run [10:07]
diana_coman: well, I lost 2 tables to random dc really... [10:11]
diana_coman: foxy'll be making some tomes atm, taking a break from all the exploring stuff [10:11]
diana_coman: also kind of curious if anyone else noticed a decrease of ords or is it just foxy showing off her more-than-bad-verging-on-horrible luck [10:12]
mircea_popescu: i got 0 ords in ... jesus, 12k bh hits [10:13]
mircea_popescu: so here i sit with my pile of sb : 503 q 222 ; 1466 q 3 973 q 207 ; 983 q 252 ; 1651 q 258 ; 2963 q 189 ; 563 q 22 and 336 q 225, doing optimal mixage calculation [10:35]
mircea_popescu: im a solid branch dj [10:36]
diana_coman: ahahaha; that's actually a very reasonable candidate for some automation on the client side (not that it's alone or anything but who knows, someone might actually do it and share it) [11:01]
mircea_popescu: i put it in #trilema as programming exercise yeah. [11:03]
mircea_popescu: should be 1st on the list for http://trilema.com/the-eulora-millenium-of-code-challenge [11:03]
lobbesbot: Title: The Eulora Millenium of Code Challenge on Trilema - A blog by Mircea Popescu. (at trilema.com) [11:03]
diana_coman: aha, that's what I was thinking: totally a very nice cs/school little project, if anyone wakes up somewhere [11:05]
mircea_popescu: it's a lot more difficult than directly apparent, ESPECIALLY if you add tailings. [11:07]
diana_coman: that's the whole story of eulora in a nutshell small enough for beginners to enthusiastically bite into [11:08]
mircea_popescu: aha. [11:09]
danielpbarron: wait if your only concern is to not lose any quality points, the solution is simple: mix odd quality with other odd quality, and even quality with even quality, both in equal amounts [12:47]
danielpbarron: i do this manually all the time [12:47]
danielpbarron: i put all the even quality stuff in the top two rows of my craft table, and the odd quality in the bottom two rows, and then I grab the smallest pile from either set, and mix in an equal amount from a different pile in the same set [12:48]
danielpbarron: then check the resulting quality, and repeat until the mixed pile is bigger than all the other piles in its quality set [12:48]
danielpbarron: eventually I end up with a mixed pile that is the same quality as an existing pile. Just mix them no matter the quantity and now there's one less pile to work with. This goes on until there are only 2 piles left: an odd and an even, usually consecutive [12:49]
danielpbarron: since storage isn't used in any of this, it goes pretty fast on the client side [12:51]
danielpbarron: it's just preparing the piles in my table that takes FOREVER [12:51]
danielpbarron: storage is infuriating [12:51]
danielpbarron: diana_coman> but actually why pairwise only? iirc you can mix up to 16 in one go using a table, no? << well I've filled up a table with different quality stuff and then moved them all to inventory at once, and I got unexpected results, but I didn't formally record any of the numbers. I just had a feeling like I lost way more quality than I had expected to [12:53]
danielpbarron: whereas the 'pairwise' method always gives me what I expected [12:53]
danielpbarron: if i mix a quality 99 with a quality 101 i will certainly get two quality 100s [12:54]
diana_coman: I do the same odd+odd and even+even, but I read the request to mean : get a desired quality without losing quality points (at least that's where I spent ages some time ago) [12:55]
danielpbarron: that's what i thought at first, but the way it was asked in #trilema made me think otherwise [12:56]
danielpbarron: when i want a certain quality, I use the same process but I set aside any resulting piles that hit the quality i was looking for. Not the most efficient way probably [12:57]
diana_coman: and optimal in the sense of fewest "steps" too [12:59]
diana_coman: hm, I guess that works as long as you have enough piles to adjust towards what you want at each step, sort of [13:01]
mircea_popescu: danielpbarron what if you don't have equal amounts ? [13:14]
mircea_popescu: ah i see. well... it's a solution, sure. the thing on #trilema is more of a general-cs thing. [13:15]
mircea_popescu: anyway : "odd and odd same qty" thing is just a narrow form. more generally, you can mix 1 part of stuff q = 0 mod 5 with 5 parts of stuff q=1 mod 5 and get an integer q also. [13:17]
mircea_popescu: and from this there's only a few short steps to modular exponentiation and then rsa. [13:17]
mircea_popescu: which is why eulora is so valuable trying to learn math/cs [13:17]
danielpbarron: diana_coman, you have a table of stuff at -351 -180 [17:34]
diana_coman: oh do I? thanks danielpbarron will check that [17:35]
diana_coman: ha, that's the first one I lost, lmao [17:36]
danielpbarron: that's not even where i found it, but where my bot dragged it to [17:37]
danielpbarron: not far off though, i caught it within a few minutes [17:37]
diana_coman: apparently eulora entered the era of lost craft tables or some such [17:39]
mircea_popescu: lol [18:55]
mircea_popescu: gah dc'd. [20:17]

Comments feed: RSS 2.0

Leave a Reply