#robotics Logs

Oct 31 2015

04:24 mrdata i have copyrighted the word obvious
04:24 mrdata you must now pay me for using it
04:54 SpeedEvil deshipu: legalities vary
04:55 SpeedEvil deshipu: In the EU, 'database right' and 'sweat of brow' copyright is recognised.
04:55 SpeedEvil Collections of facts can be copyrighted
04:56 deshipu SpeedEvil: only when in computers
04:56 deshipu SpeedEvil: and only if you copy substantial parts of the collections
04:57 SpeedEvil I don't happen to recall any 'only when in computers' part of the legislation, thoguh it's been a while since I read it.
04:57 SpeedEvil I'm currently concentrating on the welfare reform act 2013, and the personal independance payment regulations 2014
04:58 deshipu fun
04:59 deshipu I'm currently concentrating on building a keyboard
05:16 SpeedEvil http://robohub.org/blueberry-picking-robot-could-win-250000-prize/
05:18 LiohAu What is the formula to determine how many RPM are needed to achieve a given speed with a leadscrew (taking the pitch as parameter I guess)?
05:37 veverak deshipu: what you can and can't copyright
05:37 veverak is not something I would decide only on "common sense"
05:38 veverak don't believe it works this way for some time
06:18 Jak_o_Shadows LiohAu: Try looking at what some datasheets sayfor commerial leadscrews?
06:18 Jak_o_Shadows Also, it's not overy complicated
07:05 Keukpa This project is taking too long...I need more 3D printers!
07:05 SpeedEvil LiohAu: one rev of the leadscrew moves it one pitch.
07:06 SpeedEvil LiohAu: If you can't work it out from that, you should probably give up.
07:08 LiohAu SpeedEvil: The thing that is complicated is not the formula. It's just that I have to remember all the english terms that I didnot know 1 week ago (pitch for example etc..)
07:08 SpeedEvil ah
07:09 LiohAu and when you forget one, google translate does not help since they are "technical" terms
07:10 LiohAu for example, yersteday I waste 1 hour to understand the difference between rolled and ground screws :(
07:12 LiohAu also there are terms that I found in french for which it's hard to find translation. "flambage" for instance, I guess it's "buckling", but not really sure.
07:19 robopal what else do you want to know?
07:19 robopal where are you reading this?
07:36 veverak I DID IT
07:37 akem lie
07:37 veverak yaml tables for dimensions of countersunk screws, hex head screws, nuts, socket_head screws, washers, bearings and linear bearings DONE
07:39 akem mexican technology kicking your ass since 3001 BC.
08:41 robopal tables you say?
08:50 LiohAu robopal: well, my main goal now, is to have a complete list of all the options I have. If I understand correctly, there are ball screw, roller screw, acme screw, trapezoidal screw, square screw, and helicodial screw
08:51 robopal for what
08:51 LiohAu the Z translation axis for a robotic arm
08:51 LiohAu (a scara like arm)
10:06 rue_bed I uppose I shoudl write a page on ik
10:07 Tom_itx you got candy for me rue_bed?????
10:07 Tom_itx i'm gonna egg your house if you don't!
10:07 rue_bed tonight
10:07 rue_bed havre to carve the pumpkin still!
10:07 rue_bed robotics in 2ish hours
10:08 Tom_itx blinky led for the pumpkinhead
10:09 rue_bed meh, no time
13:54 veverak hmm
13:54 veverak is it safe to assume
13:54 veverak that all "micro" servos share all dimensions?
13:54 veverak in a way that when I design 3d part to house servo it will match to any servo of same "size"
13:55 Snert ummmm, no, unfortunately.
13:56 Snert rc helis taught me that. Hole pattern spacing, mounting ears, and even case dimensions will just vary. They are pretty close, but they vary.
13:57 Snert I've found that sticking with 1 brand, such as Hi-Tec will stop the madness.
13:58 veverak that's what I was afraid of
13:58 Snert and the cheaper you go, the more variance.
13:58 veverak hmm
13:58 veverak so, realisticaly
13:58 veverak making dimensions for "towerpro micro servo"
13:58 veverak "hitec normal servo"
13:58 veverak etc... could work?
13:59 Snert especially with these totallt crap servos you see in these arduino tutorials. The "tower" brand.
13:59 veverak yeah
13:59 veverak they are
13:59 veverak but for some stuff, it's good enough
13:59 Snert it's hard to spend less than those servos.
14:00 Snert hitec does a line of micro servos. yes...and if you stick with that exact servo you'll be fine.
14:00 Snert or....
14:00 Snert make your mounting design accomodate slight variations.
14:01 veverak yeah
14:01 Snert they sell servo adapter plates to help with that made out of carbon fiber and sometimes aluminum.
14:01 Snert but fuckthat.
14:03 veverak well
14:03 veverak practically
14:03 veverak for me it's just changing dimensions and script generates part for different servo
14:03 verak goes full parametrization usu
14:04 Snert hobbyking has a good assortment of servos where you can compare case/mounting dimensions and perhaps get an idea of how much variance within a given class that you might encounter.
14:06 Snert but they have not been accurate 100% of the time either.
14:06 veverak and they only got the width/height/depht
14:06 Snert yea.
14:06 Snert that's why I say HiTec or Futaba or perhaps JR where you can get a full specsheet on the manu website.
14:07 verak didn't found on h
14:07 veverak http://etechpk.net/wp-content/uploads/2014/10/113-4.jpg
14:07 veverak also, this saves mi for this project
14:07 veverak :)
14:08 veverak *saves me
14:08 Snert another big area of variance is distance from the case top up through the top gears and control arm.
14:09 Snert I cann that the "tower height".
14:09 Snert call
14:10 Snert seems like I always have to file holes bigger, space the servo in or out with washers and stupid stuff like that.
14:10 Snert or mounting hole patterns are different.
14:11 Snert some use 4, some use 2.
14:12 Snert http://hitecrcd.com/files/2013_HRU_ServoMat_2.pdf
14:13 Snert prolly doesn't help though.
14:16 Keukpa this is a good website: http://www.servodatabase.com/servos/all
14:37 LiohAu what is the "start" on a lead screw?
14:38 LiohAu I thought that the pitch was the distance between two "steps", but it seems that it's also the definition for "lead", and depending on the "start" the "lead" definition change. But I don't know what is the start :(
14:43 LiohAu oh ok I just understood what the start is, so basically the pitch is the distance between two groove of the thread, while the leads use the number of starts..
14:45 LiohAu And I guess that there aren't weird screw having starts count != end count, right?
19:07 veverak Keukpa: not really
19:07 veverak width/depth/height for servo dimmensions is not enough
19:32 superlou So, I sucessfully made a numerical solution to the direct kinematics in my robot
19:32 superlou Unfortunately, the IK is now hard. I'm trying to use a particle swarm method, but it just seems unstable, and will be really slow if it does work
19:33 superlou does anyone have any tips for numerical IK solvers that don't have analytic direct kinematics?
19:38 rue_house WHY CANT YOU PEOPLE DO THIS
19:38 rue_house what kinda robot IS it!?
19:38 superlou rue_house, https://vimeo.com/136165642
19:38 superlou That kind
19:39 superlou Yeah, kinda a vague problem statement, sorry. More looking to start a conversation.
19:39 superlou I've only ever done IK (back in college) for stuff with analytical DK
19:39 rue_house well, its not a normal 1d axis, so ok
19:39 rue_house its still not hard tho
19:40 rue_house your defining your origins wrong
19:40 superlou what do you mean?
19:40 rue_house its two simple calcs
19:40 rue_house is it suspended or on a pivot?
19:41 rue_house oh its suspended
19:41 rue_house ok, getting mor complex, but not htere yet
19:41 superlou Not sure I follow. The platform (white portion) is positioned by 3 parallel actuators
19:41 superlou then the eyes are attacehd to one more servo mounted in the center
19:41 rue_house yup
19:41 rue_house I see now
19:41 superlou So the only way I could figure out how to solve for the head position based on servo arm angles was to use a numerical solver
19:42 rue_house heh
19:42 superlou Then the position of the eyes relative to that are pretty straightforward
19:42 rue_house your doing forward or reverse kinamatics?
19:43 superlou I've done forward so far https://github.com/superlou/deskbot/blob/master/kinematics/simple/dk_model.py
19:43 superlou And I was hoping to use that to do inverse via particle swarm, or cyclic descent or something
19:43 superlou But i'd like to do it in realtime, which is proving to be challenging
19:43 superlou (let alone solving)
19:44 superlou I was htinking about a neural network approach, since I can run the forward kinematics really fast to come up with training data
19:44 superlou but the last time I tried doing that, the neural net had issues due to the angles being able to roll-over
19:44 superlou And i couldn't figure out how to constrain it
19:45 adam789654123 what problem is it supposed to solve exactly?
19:45 adam789654123 the neural net i mean
19:45 adam789654123 im not farmiliar with reverse kinamatics
19:46 rue_house what are you controlling it with?
19:46 superlou Given the height of the eyes point of rotation, and the unit vector the eyes are looking towards, it should solve for the angle of 4 servos
19:46 superlou rue_house, right now, python on a raspberry pi, though if i get it basically working, I'll probably move to C for speed
19:46 rue_house yea
19:47 superlou it's just faster to play around in python for now, and I've rigged up a simple viewer to examine the geometry of the model using pyqtgraph
19:48 rue_house arg, I'd like to get into it now but I ahve to dot tings
19:48 superlou no problem :)
19:48 superlou It's actually turned out to be a much more interesting problem than i thought it would be
19:48 superlou I just wanted to make a box that could open and close itself, and the kinematics have proven to be really challenging
19:51 adam789654123 "angle" of 4 servos or "angles" of 4 servos?
19:51 superlou adam789654123, 1 angle per servo
19:51 superlou 3 in the base, 1 in the white platform
19:51 adam789654123 "solve for"
19:51 adam789654123 i guess you need to solve considering the geometry involved?
19:52 adam789654123 fyi, im working from the terminal
19:52 superlou direct kinematics: given 4 servo angles, determine unit vector the eyes are looking and hte height of the platform
19:52 adam789654123 i cant look at your robot
19:52 superlou ah, ok.
19:52 superlou direct kinematics I solved numerically with a solver
19:53 adam789654123 "unit vector"
19:53 superlou essentially the robot has 3 servos in the base that move each corner of a triangular platform up and down to control it's tilt and vertical position. A 4th servo in the platform itself rotates the eyes around
19:53 superlou adam789654123, vector of lenght 1. The eyes can't look "harder" or "less hard" :) Only in a direction.
19:53 adam789654123 oh
19:53 superlou The eyes are attached to the 4th servo sitting horitzontally in the platform
19:54 adam789654123 you want the eyes to respond as the thing moves?
19:54 adam789654123 to sort of keep them level or something?
19:54 superlou adam789654123, down the road, that's the idea, but for now i just care about hte kinematics
19:55 adam789654123 how does that constrain the problem?
19:55 adam789654123 caring only about the "kinematics"?
19:55 superlou So in my current model, for 4 servo angles, I can predict where the eyes are looking and how high they are.
19:55 superlou I need to do the opposite.
19:55 adam789654123 where the eyes are and how high can determine the angle of the motors?
19:56 superlou yeah
19:56 adam789654123 is that "reverse" kinematics?
19:56 superlou Think of it like controlling your hand. You decide where you want your finger pointing, but you don't independently think of each bone's position to accomplish that
19:56 superlou Yeah. The term is traditional "inverse kinematics"
19:56 adam789654123 ah
19:56 adam789654123 ok
19:57 superlou Whereas if you were shifting the bones around, and then measuring where your finger ends up, that's "direct kinematics"
19:57 adam789654123 im thinking about that
19:58 anonnumberanon I'll have to do that with my biped robot.
19:58 adam789654123 i dont know the angular relationships of my skeloton to point at the door knob
19:59 adam789654123 thats inverse kinematics?
19:59 superlou adam789654123, yep
19:59 adam789654123 i just know finger pointing at the doorknob, and the body solves the geometry
20:00 superlou IK solvers are a huge research topic in themselves.
20:00 adam789654123 direct kinematics starts with the angular geometry, and then thinks, where is my finger pointing?
20:00 adam789654123 ok
20:00 superlou adam789654123, you got it.
20:01 adam789654123 thats interesting
20:01 adam789654123 about the solvers btw
20:01 superlou One of the more interesting/tricky things is that there's typically many IK solutions for the same inputs
20:01 adam789654123 hmmm....
20:01 superlou There's many ways to orient the bones of your arm that achieve the same end result
20:01 adam789654123 well... im quite fascinated in geometry
20:01 anonnumberanon What is the best way?
20:01 adam789654123 so much so that ive read Euclids Elements
20:02 adam789654123 and im studying algebra now so that i can get into modern approches
20:02 adam789654123 computational geometry
20:02 anonnumberanon If trying to find the best way in real time, that must be a lot of calculations.
20:02 superlou anonnumberanon, there's no real "best" way. There's a ton of different techniques, where some are better for some problems than others.
20:03 adam789654123 "real time" itself could be relative
20:03 adam789654123 for example, frames per X
20:03 superlou http://www.cescg.org/CESCG-2002/LBarinka/paper.pdf is a pretty good walkthrough on the terminology and some appraoches
20:03 adam789654123 great
20:03 adam789654123 no pdf reader yet
20:03 adam789654123 but its going in my robotics file
20:03 anonnumberanon By real time I mean, we are in position x, and want to move the finger to position y, if we were in position z instead the calculations would be different.
20:04 anonnumberanon I imagine maps of power efficient moves here.
20:04 superlou anonnumberanon, yep. Also, there may be obstacles your finger can't move through along the way
20:04 anonnumberanon Hash maps. AI typed search algorithms.
20:04 superlou So you combine a pathfinding algorithm with your IK solver
20:04 adam789654123 interesting
20:05 adam789654123 i guess you could consider it, from the robot's point of view, information gathering through the act of performing the procedure
20:05 anonnumberanon I have 17 servos and I think my controller circuit will sense all the currents going through each of them. Then start playing with different moves and see which ones require more energy.
20:05 adam789654123 whereby a map is being constructed as the operation is being performed
20:06 superlou anonnumberanon, there are learning algorithms like that, but you have to be really aware of the noise in the measurements for a learning algoritm like that
20:06 superlou adam789654123, at this point, i'm considering a map/interpolation brute force approach, but that's boring
20:06 superlou neural net is my next attemp
20:06 anonnumberanon By map i mean a collection of possible solutions for a certain move, depending on starting positions (then the accelerometer feedback adds a whole world of hurt to this).
20:07 adam789654123 well.... so much worth considering here
20:07 superlou Do you have a specific project in mind?
20:07 anonnumberanon me or adam?
20:07 superlou DK and IK is definitely one of those things you want to discuss with a specific geometry in mind
20:07 superlou Either? :)
20:07 superlou I'm doubly boned since my direct kinematics aren't solved analytically and require a numeric solution
20:08 adam789654123 i was considering a solver
20:08 adam789654123 like a deduction engine
20:08 adam789654123 but the problem is setting it up properly
20:08 anonnumberanon I want to keep my humanoid biped robot standing, first in probably the least power consuming position, but keeping the humanoid robot's style in mind (can't just put it on four legs that would be stupid).
20:09 superlou anonnumberanon, does it need to walk, or just stand?
20:09 anonnumberanon superlou, you're getting ahead there haha, but it needs to fight eventually, other bipeds.
20:10 superlou if you just need to stand, you may be able to do a relatively straightforard balance thing with an acceleroamter
20:10 anonnumberanon If i say what the need is for the final spec it is really scary.
20:10 superlou anonnumberanon, that's pretty ambitous :)
20:10 superlou does it need to have dynamic stablity while moving, or is static stability ok?
20:11 superlou http://www.cs.cmu.edu/~cga/legs/stability.pdf is pretty neat
20:13 anonnumberanon I don't know too much about robot walking, all i know is the Asimo and the robots it will be fighting, which have a lot more mechanical solutions than electronic solutions that make them walk and not fall down. Not sure the difference between dynamic and static stability. Dynamic is using the accelerometer? Feedback control? In any case what I envision is making it stand, then hooking up the accelerometer so that if I push a
20:13 anonnumberanon robot a bit, it will move its body or EVEN take a step so that it does not fall. That seems like a pretty good first challenge to get things going.
20:14 superlou anonnumberanon, there's different terms for it in different papers, but in my experience static walking means the center of gravity is always over the foot that is on the ground. Like if you mechanically step slowly, always keeping your "weight" on the foot that is not moving.
20:14 adam789654123 wonder how you would make a balance sensor?
20:14 superlou Dynamic is like walking where if you were to stop mid stride you would fall
20:15 superlou adam789654123, Typically, you could use accelerometers and gyros to monitor the center of mass of the robot's body
20:15 superlou Then based on the servo positions, you have a model of your robot's geometery and masses
20:15 adam789654123 i was thinking something different
20:15 adam789654123 much simpler
20:15 superlou from there you can determine if your center of mass is over a foot, and if you are going to fall
20:15 superlou If it was easy, you'd see a lot more walking robots :)
20:16 adam789654123 heres a great idea (imo)
20:16 adam789654123 you have an insulated cup, and two nodes at each side sticking part way into the cup
20:16 superlou note: as soon as mass comes into it, you're talking "dynamics" rather then "kinematics"
20:16 anonnumberanon Heh, so far I've ordered parts for the servos' power supply, made a quadcopter so I at least know a bit about the gyro/accelerometer, and I've moved one of the servos with a joystick, I've come to the conclusion the servos like 5V, struggle as I raise the voltage and act completely erratically at 7V and up. They also take about 300 miliamps without torque applied, and go up to 1 Amp when I'm holding the shaft rather strongly and
20:16 anonnumberanon command the servo to spin. The buck converter regulator is coming in the mail Monday, so I can power more than just one servo (any voltage, adjustable, and max power at 14 amps).
20:17 adam789654123 the cup is filled with a conducting fluid
20:17 adam789654123 when you tilt the cup, the connection breaks
20:17 anonnumberanon
20:17 superlou adam789654123, you're describing a fluid accelerometer. We use them in older aviation stuff
20:17 adam789654123 ok
20:17 superlou Mostly moved to MEMS stuff at this point, but they are cool when you come across them
20:17 adam789654123 something with resistance might be better
20:18 adam789654123 with a relative resistance
20:18 anonnumberanon I get it now. I have been thinking about this longly. No idea which method will be best for power consumption, and walk stability and walk speed.
20:18 superlou adam789654123, just use an IC :)
20:18 adam789654123 depending on the level
20:18 adam789654123 is it certainly better
20:18 adam789654123 ?
20:18 anonnumberanon superlou, let me show you a video of one of the tests the robot must pass before being allowed to fight.
20:18 superlou anonnumberanon, you shouldn't raise the servos voltage above their ratings. if you need more torque, you need a different servo
20:18 adam789654123 i like to let the air dry my wet dishes
20:18 adam789654123 i didnt have to invent anything for that
20:19 superlou adam789654123, usually it's more predictable to use the MEMS rather than the fluid parts. They are also a bit fragile
20:19 anonnumberanon superlou, they have a LOT of torque at 5V, I was just mapping their voltage rating.
20:19 adam789654123 MEMS...
20:19 adam789654123 ill have to look that up
20:19 superlou adam789654123, micro electro mechanical system
20:19 anonnumberanon here it is: https://www.youtube.com/watch?v=rNlC7RYMuho
20:19 adam789654123 ok
20:20 anonnumberanon I am terrified by the work needed to accomplish this hehe.
20:20 superlou adam789654123, Really small springs and masses made out of silicon in a chip. It's how your phone knows it's orientation
20:20 superlou anonnumberanon, checking it out...
20:20 adam789654123 ic
20:20 anonnumberanon superlou, also I can't really figure out if that is static or dynamic, dynamic I think would surely add great speed to the walk/run of the robot.
20:20 anonnumberanon (like humans?)
20:21 superlou anonnumberanon, with that video, it was hard to tell. That was an impressively fast moving one
20:21 adam789654123 if it was of substatial size, it *might* not matter really
20:21 adam789654123 the robot i mean
20:21 adam789654123 those parts are likely used because of the application
20:21 superlou trying to find a good video on static vs dynamic
20:21 adam789654123 even though i do find it interesting
20:22 adam789654123 i like to see if i can come up with something very simple
20:22 adam789654123 and easily overlooked
20:22 superlou adam789654123, your fluid idea is solid, but it only reports a few different positions. the accelerometer IC gives you measurements in multiple dimensions
20:22 superlou Plus, the fluid has a significant mass, which can cause it to lag, or oscilate making fast measurements difficult
20:22 adam789654123 i see
20:24 anonnumberanon what is your project adam789654123 ?
20:24 adam789654123 now?
20:24 adam789654123 a function that im trying to factor
20:24 anonnumberanon or ever..
20:24 adam789654123 almont a hundred lines of code
20:24 adam789654123 never built a robot
20:24 anonnumberanon what does factor mean?
20:25 adam789654123 im using an analogy to algebra
20:25 adam789654123 make it smaller
20:25 adam789654123 find the peices that make it up
20:25 anonnumberanon or, "more atomic"
20:26 superlou adam789654123, in software, it's often called "refactoring": the process of taking code that works and cleaning it up to be more elegant
20:26 anonnumberanon are you trying to make the function faster or something?
20:26 adam789654123 superlou: yeah, i like the term "factor" better
20:26 superlou fair
20:26 adam789654123 its more specific to my frame of mind
20:27 adam789654123 "refactor" implies that its is factoring again
20:27 adam789654123 which is kind of funny
20:27 adam789654123 because the codebase probably changed since the last time it was factored
20:27 adam789654123 so....
20:27 adam789654123 i like "factor"
20:27 anonnumberanon It is funny that there is actually an entire book dedicated to refactoring code..
20:28 adam789654123 many of them
20:28 adam789654123 :)
20:28 anonnumberanon what language is it in?
20:28 adam789654123 common lisp
20:28 adam789654123 its a programmable interface
20:28 adam789654123 and so far, i REALLY like it
20:28 anonnumberanon Ah, so you can start by porting it to C ;)
20:28 adam789654123 but i need to move things around
20:28 adam789654123 haha
20:29 adam789654123 well... i am using Clisp
20:29 adam789654123 now that you mention it
20:29 adam789654123 ;)
20:29 anonnumberanon intriguing
20:29 adam789654123 it took awhile to actually get this thing together
20:29 adam789654123 alot of thinking
20:29 anonnumberanon superlou, for the record I didn't go above 7v on that servo test heh
20:30 adam789654123 almost been working on it a week
20:30 anonnumberanon adam789654123, what does it do?
20:30 adam789654123 but the core ideas are quite beautiful
20:30 adam789654123 do you know what recursion is?
20:30 anonnumberanon Yes.
20:30 anonnumberanon I have to do some soon in my C class.
20:30 adam789654123 imagine have a read within each recursive call
20:31 superlou if it wasn't for the cost of parenthesis, i'd like lisp better :) Anyway, time for me to pass out. have a good night!
20:31 adam789654123 gnight superlou
20:31 adam789654123 :)
20:31 anonnumberanon I kind of dread it too because I hated it in Java but maybe because I didn't apply myself to it enough.
20:31 adam789654123 its about the programming for me
20:31 adam789654123 besides, what other language has this as a data strucuter: (Y)
20:31 adam789654123 :P
20:31 adam789654123 anyhoo
20:31 anonnumberanon don't know..
20:32 adam789654123 there is a read on each recursion
20:32 adam789654123 and an element that is passed down on each recursion
20:32 adam789654123 there are functions that operate on that "object"
20:32 anonnumberanon adam789654123, have you read SICP?
20:32 adam789654123 i want to
20:32 adam789654123 its on my reading list
20:32 adam789654123 looks great
20:33 adam789654123 you/
20:33 adam789654123 ?
20:33 anonnumberanon go, run like you've never run before, read it
20:33 adam789654123 ok
20:33 adam789654123 :D
20:33 anonnumberanon no I haven't but I hang out in a channel where all the users swear by it
20:33 adam789654123 so you would understand some pseudocode
20:33 adam789654123 ok
20:33 adam789654123 fun (element functions)
20:34 anonnumberanon they are programmers and I am more of an electrical engineer so I only code in C, but I'm open to looking up more things
20:34 adam789654123 is a function that takes in an element, and the functions that operate on that element
20:34 anonnumberanon *out of time*
20:34 adam789654123 you understand anonnumberanon ?
20:34 adam789654123 fun (element functions)
20:34 adam789654123 is a function that takes in an element, and the functions that operate on that element
20:35 anonnumberanon kind of, in C it would be sending as an argument th epointer to a function or the pointer to a list of functions..
20:35 adam789654123 its like a function signature
20:35 adam789654123 not the entire thing
20:35 adam789654123 but just to give an idea of the function interface
20:35 adam789654123 lets say this function calls itself
20:36 adam789654123 lets say that first this function reads in which function to use to transform the element
20:36 adam789654123 then it calls itself with the transformed element and the same functions
20:36 adam789654123 in this way, you could have a trivial UI to operate on a specific kind of object
20:37 adam789654123 my project is an extension of this idea
20:37 adam789654123 maybe kind of confusing anonnumberanon ?
20:37 adam789654123 np
20:37 anonnumberanon it is
20:37 anonnumberanon very
20:38 adam789654123 we take stuff for granted all the time
20:38 adam789654123 ive been doing this stuff for almost a decade
20:38 adam789654123 :)
20:38 anonnumberanon wow
20:38 anonnumberanon it must be like an art for you now
20:38 adam789654123 for me personally, yeah
20:38 adam789654123 its a form of expression
20:39 adam789654123 im working on a trivial CAS
20:39 adam789654123 or computer algebra system
20:39 adam789654123 i just finished something i call an "recursive interupt interface" or "rii"
20:39 adam789654123 whereby you can program interupts in a very flexible robust way
20:40 adam789654123 but i need an interface for the interupts to be of any great use
20:40 anonnumberanon in x86?
20:40 adam789654123 yeah
20:40 anonnumberanon hardware interrupts or something?
20:41 adam789654123 oh... in the language
20:41 adam789654123 so i can write a function then register it for an interupt pattern
20:42 adam789654123 if any of the interupt patterns evaluates to true, the evaluation is aborted
20:42 adam789654123 its ano
20:42 adam789654123 its another neat device
20:42 adam789654123 here is the simplest idea to get it
20:42 adam789654123 if i have a function that calls itself
20:43 adam789654123 f1 > f1 > f1 ....
20:43 adam789654123 how can i write code to test for and break the cycle?
20:43 adam789654123 the pattern is quite simple
20:43 adam789654123 you have another functions sit between each function call
20:44 anonnumberanon I see.
20:44 adam789654123 instead of having f1's calls to itself actually call itself, they are calls to function g requesting a call to f1
20:44 Snert 2 identical functions that differ only in name. One calls the other.
20:44 adam789654123 actually, g calls f1 after checking the conditions
20:45 adam789654123 f1 actually *cant* call itself
20:45 adam789654123 its calls to itself are actually a call to g
20:45 adam789654123 f1 > g > f1 > g ...
20:45 adam789654123 thats the meaty simple core idea that you can build around
20:48 adam789654123 anonnumberanon: which channel is do they reccoment that book?
20:49 anonnumberanon pm
20:49 adam789654123 i messed it up
20:49 adam789654123 sorry
20:49 adam789654123 can you do it again
20:50 anonnumberanon come say hi sometimes, they ahve a few really smart CS people in there who don't take themselves too seriously
20:51 adam789654123 cool
20:51 adam789654123 thanks again
20:51 anonnumberanon np
20:51 adam789654123 :)
20:53 anonnumberanon we got a lot of bots in there that do cool stuff
20:54 adam789654123 oh
20:54 adam789654123 awsome
23:57 Eziodjinn Hey guys and girls I have a question about the Roboard 110, are there any variants of this or alternatives I can use that aren't so pricey? I'm trying to build a quadruped robot, the original build used a RB110 but this was a few years ago and I'm wondering if something like an arduino setup would be comprable?
23:59 Eziodjinn Or maybe even like a raspberry pi, it needs to run 8 servos and do some movement gaits on a small bot though, about a foot long