#robotics Logs
Oct 18 2017
#robotics Calendar
12:10 AM evilroot: Planned out in advance
12:13 AM evilroot: So as to give an excuse for all that Kickstarter cash they blew
12:13 AM evilroot: Pissed me off so much
12:18 AM rue_: zlog
12:23 AM evilroot: Give me even a damned light duty rental excavator and I would curb stomp both "giant robots" in under 30 seconds
12:23 AM evilroot: I mean that's sad
12:37 AM rue_: :( I didn't take pics of my milling machine bearing rebuild
12:37 AM rue_: nick rue_mohr
12:43 AM rue_ is now known as rue_mohr
12:43 AM rue_mohr: with time, I'll get it right
03:35 AM z64556 is now known as z64555
03:59 AM edmoore_ is now known as edmoore
07:54 AM justasking: if I want my robot to know about the location of some moving objects, what are the options I have?
07:55 AM justasking: like what kind of stuff can I attach to them (and which counterpart to my robot)
08:07 AM deshipu: justasking: more details needed
08:08 AM deshipu: justasking: range, accuracy, speed, budget
08:09 AM deshipu: I think the most common approach is to use fiducials and a camera
08:16 AM SpeedEvil: I recommed an array of mics, and every device playing loudly various different covers of The Macarena.
08:18 AM deshipu: backwards
08:23 AM deshipu: https://www.youtube.com/watch?v=keb0B11zj5g
08:29 AM deshipu: but if the environment is pretty drab, you might even get away with just detecting color blobs
08:33 AM SpeedEvil: Or flashing LEDs
08:33 AM SpeedEvil: Retroreflectors can give you massive gains, compared with 'white'
08:34 AM SpeedEvil: qrcodes, painted on retroreflectors, with selective illumination can pop out startlingly
08:38 AM deshipu: april tags, qrcodes are too hard to find
08:53 AM SpeedEvil: Hence the retroreflector, which pops out at 10* white brightness
08:55 AM justasking: but with a camera it is pretty hard to determine at which distance the objects are or not?
08:55 AM justasking: or you'd need some good computing power at least
08:55 AM deshipu: justasking: no, the size is proportional to the distance
08:56 AM justasking: linearly?
08:57 AM deshipu: depends on your lens
08:57 AM deshipu: there is a formula
08:57 AM justasking: ah ok, thank you
08:57 AM deshipu: but if you remember how perspective works, then you will see that it has to be pretty much linear
08:58 AM justasking: but I'd have to scan the whole image in a step, I thought more about sensors which directly give me some information
08:58 AM justasking: I want my robot to get some small objects, e.g. tennis balls as fast as possible
08:58 AM justasking: like a poor mans dog
08:59 AM deshipu: define "directly"
08:59 AM SpeedEvil: justasking: locating objects is a hard and not generally solved problem. What's your budget. Active tags reporting their position can help
08:59 AM deshipu: that openmv camera from the video pretty much does all the work for you
08:59 AM justasking: yeah I mean with accelaration sensors I should be able to compute the position all the time if I know the starting position of the object I am throwing
08:59 AM SpeedEvil: justasking: yeah - no
08:59 AM justasking: ah I did not check the video I thought it was a joke :D
08:59 AM justasking: wait a second
09:00 AM SpeedEvil: justasking: accelleration sensors _rapidly_ lead to a drift in position, due to the triple integration you need, never mind angular uncertainties
09:01 AM justasking: ahh no that is to expensive and also to advanced for my needs I guess
09:01 AM deshipu: yet what you describe is even more advanced and expensive
09:02 AM deshipu: it's one of those problems that seem like they should be simple, but really are not
09:02 AM justasking: SpeedEvil, ah ok, yeah maybe it is harder than I thought
09:02 AM justasking: ah and position is hard since there may be sensors reporting the distance to each other but I'd need more information right?
09:03 AM justasking: something like gps with multiple distances
09:03 AM justasking: maybe a simple camera really is my best bet
09:03 AM justasking: and I'll color the table tennis balls in some bright color
09:03 AM deshipu: a camera on the ceiling and fiducials on all objects you want to track would work best
09:03 AM justasking: I'd like to have the camera on my robot
09:04 AM deshipu: justasking: can you put a blinking light on the ball?
09:04 AM deshipu: justasking: or inside the ball
09:04 AM deshipu: it could even blink in infrared
09:04 AM deshipu: then it's much easier to locate on the image
09:05 AM justasking: ah yes thank you
09:05 AM justasking: one more question, do you know a good retailer for small electronics stuff in europe/germany?
09:05 AM deshipu: tme.eu
09:07 AM deshipu: I use mouser too, for stuff tme doesn't have
09:31 AM justasking: tme doesn't have wheels that I can directly stick to servos
09:32 AM justasking: https://shop.pimoroni.de/products/pololu-wheel-90x10mm-pair
09:32 AM justasking: this kind of stuff
09:33 AM justasking: ah and can I use the raspberry pi camera module with a beaglebone or an arduino?
09:33 AM SpeedEvil: justasking: No
09:34 AM justasking: so each board has its own cameras which are only compatible with that board? how come?
09:34 AM SpeedEvil: justasking: Because there is no incentive to standardise.
09:34 AM SpeedEvil: And different cameras work better for different things.
09:34 AM deshipu: justasking: wheels are not electronics
09:35 AM justasking: yes I am not sure the raspberry pi camera module is what I need, since for my task I don't need very high resolution images
09:35 AM deshipu: justasking: this kind of stuff I buy at rc model shops: https://abc-rc.pl/
09:35 AM justasking: lower resolution may be better since I probably can work faster with them
09:35 AM deshipu: justasking: you can always switch the camera to a lower resolution
09:36 AM justasking: also on hardware side? I mean doing it on software side would make a lot of work
09:37 AM deshipu: this will be a lot of work anyways
09:37 AM deshipu: opencv will probably help you
09:39 AM SpeedEvil: Also if you're a cheap bastard: http://www.ebay.co.uk/itm/Mini-3D-Photograph-Stereoscopic-Camera-Lens-w-Clip-For-iPhone-Smart-Phone-/262891018187?hash=item3d358697cb:g:HlcAAOSwCU1YxEzt
09:39 AM SpeedEvil: + raspberry pi camera will give you some modicum of 3d
09:39 AM deshipu: oh, I had successes with cheap laptop camera modules
09:39 AM SpeedEvil: Out to several meters.
09:40 AM deshipu: https://hackaday.io/project/3400-toshiba-8071a2gb-webcam-module
09:40 AM SpeedEvil: It's nicer in some ways than two cameras, as it is synced optically of course.
09:40 AM deshipu: they are 3.3V, but work with usb
09:41 AM deshipu: you can get them really cheap from junked laptops
09:41 AM deshipu: you just need to figure out which wire is which, but that's not to hard
09:41 AM justasking: wait I can just use camera modules of an old phone or laptop? are there any tutorials on how to do that?
09:41 AM SpeedEvil: justasking: no
09:42 AM SpeedEvil: justasking: he is speaing specifically of USB camera modules.
09:42 AM SpeedEvil: There is no simple way of using random camera modules from phones , and laptops only if they have internal USB interfaces.
09:42 AM SpeedEvil: (which doesn't help you at all for arduino)
09:43 AM justasking: but it would help me if I'd use a raspberry pi or beaglebone? so do I understand it right that they are just plugged into an internal usb and I could plug that into my usb? how do I find out if the lenovo thinkpads t4xx do that?
09:43 AM deshipu: most old laptops have usb modules
09:43 AM deshipu: yes, basically
09:44 AM SpeedEvil: It doesn't help you any more than using an off-the-shelf webcam
09:44 AM SpeedEvil: they're basically identical.
09:44 AM SpeedEvil: Thoguh may of course be smaller
09:44 AM deshipu: justasking: well, the eaesiest way would be to run lsusb and see if camera appears in the list
09:45 AM justasking: SpeedEvil, well it helps me in the way that I would not have to buy one or not
09:45 AM justasking: and I'd like to get to know about stuff like this
09:46 AM SpeedEvil: As a general point, going for scrounged stuff that is never going to be available again can really, really slow you down.
09:46 AM justasking: like even if I don't save much money I would just do it out of interest if I could dissassemble my broken laptop and use parts
09:47 AM deshipu: right, it makes your project impossible to repeat
09:48 AM justasking: well all I need is to buy another lenovo laptop :D
09:48 AM deshipu: you can buy parts
09:48 AM justasking: yeah no you are right but I think I want to try to do that
09:48 AM deshipu: for a while at least
09:48 AM SpeedEvil: They can change internal parts without notice even on the same model.
09:48 AM deshipu: then just get a broken usb cable
09:48 AM justasking: I mean even if it doesn't work, I'll just buy a camera then and do it with a camera
09:48 AM justasking: no it was a joke, I would not buy a new laptop for a new camera :D
09:49 AM deshipu: getting them from junk is best
09:49 AM deshipu: I got my for $2
09:49 AM deshipu: mine
09:50 AM deshipu: best to get a few identical ones, then you can really experiment without being afraid of breaking them
09:50 AM deshipu: you can often get them cheap at ebay too
09:54 AM rue_mohr: I had a horrid time with a new lonovo laptop and will never go there again
09:55 AM rue_mohr: bad motherboards, 3 of them, in a row
09:55 AM rue_mohr: and I had to fight to not have to pay shipping of the laptop to them
09:55 AM justasking: I really like the thinkpads from earlier
09:55 AM justasking: the t4x0 ones are just great
09:56 AM rue_mohr: yea, back when ibm made them it was fine
09:56 AM rue_mohr: after lonovo took over it went to crap
09:56 AM justasking: mine is already made by lenovo I think its fine
09:56 AM justasking: or at least better than any other notebook I had
09:56 AM justasking: it is pretty solid and just does everything I want, doesn't break and has good bios legacy support for linux
09:57 AM rue_mohr: had to chip it back _3_ time in a row
09:58 AM rue_mohr: aka, got it back, wouldn't start, motherbaord failure, send it back
09:58 AM rue_mohr: back to lonovo, the seller wouldn't take it
10:02 AM edmoore: i still have my x201 around
10:02 AM edmoore: i use a macbook pro for my main machine, it's before any of the touchbar stuff, it's still grand but i don't know if i'll have another apple when this one gives up
10:27 AM Snert: edmoore: you got an identical twin brother?
10:28 AM edmoore: nope
10:28 AM edmoore: why do you ask?
10:28 AM Snert: you're not the same ed moore that I knew is all.
10:29 AM edmoore: how do you know
10:29 AM edmoore: and how many identical twins are given the same christian name
10:29 AM Snert: cuz you'd have an identical twin brother if you were.
10:30 AM edmoore: i'm not sure that follows
10:42 AM Snert: identical twins having the same name is all you. The identical twins I knew did not have the same name.
10:42 AM Snert: and I've never seen identical twins with the same name.
10:43 AM edmoore: no exactly
10:44 AM edmoore: so if i'm not the same ed moore you knew, i.e. you knew an ed moore but it's not this ed moore, then it's unlikely that we are each other's identical twin, because it's unlikely that indentical twins have the same name
10:45 AM Snert: Ed and Don Moore were their names. Couldn't tell them apart if they were side-by-side.
10:45 AM edmoore: i see!
10:45 AM edmoore: i don't know a don moore
10:46 AM edmoore: it's ringing a bell though
10:46 AM edmoore: maybe in the space undustry
10:46 AM edmoore: industry*
12:00 PM veverak: aaaand here we go
12:00 PM veverak: rosfarm building debian jessie armhf
12:00 PM veverak: :)
12:43 PM weyland|yutani: veverak, is a farmer now building a new hen house
12:49 PM veverak: :)
12:49 PM veverak: more or less
12:49 PM veverak: yeah
01:24 PM climbing_up: Are there really no good 2D LIDARs between $350 and $5000 that work outdoors? It looks like you either spend $350 on the Scanse Sweep or $5000+ on a Hokuyo unit
01:30 PM weyland|yutani: climbing_up, whats the problem outdoors?
01:30 PM weyland|yutani: range?
01:31 PM climbing_up: Range
01:31 PM deshipu: that big hot source of radiation in the sky?
01:32 PM climbing_up: I'm pretty sure that's where the problem comes from, yes
01:33 PM weyland|yutani: deshipu, that thing is a lie and you know it
01:34 PM weyland|yutani: climbing_up, i guess you need a resolution of a few centimeter for what 60 meter? or moar?
01:36 PM deshipu: yeah, nobody in their right mind would make a world orbit a huge nuclear furnace
01:36 PM deshipu: that would be stupid
01:36 PM weyland|yutani: in a vacuum...
01:37 PM deshipu: a sane person would go with the traditional elephants and turtles
01:37 PM climbing_up: If I could get reliable 15 meter performance with 10 centimeter resolution I would be thrilled
01:38 PM climbing_up: There's the LIDAR-Lite v3 that claims to have 40 meter range, but it doesn't spin... https://www.sparkfun.com/products/14032
01:40 PM weyland|yutani: climbing_up, i just read its specification
01:40 PM weyland|yutani: its garbage
01:40 PM weyland|yutani: >10 cm @ 5 meters
01:42 PM climbing_up: How much greater than 10 cm? :) I might be able to work with that. I'm trying to detect cars
01:44 PM weyland|yutani: climbing_up, i dont know
01:44 PM weyland|yutani: climbing_up, but i dont think its 10 or close to 10
01:45 PM climbing_up: Oh I see that on the spec sheet now
01:46 PM weyland|yutani: oh cars
01:46 PM climbing_up: Maybe I overestimated the resolution I need
01:46 PM weyland|yutani: i was more thinking of small obstacles
01:57 PM climbing_up: I could probably get away with even 40 cm or 50 cm resolution as long as I was able to detect 5 to 10 points in a line and recognize it as a car
02:00 PM climbing_up: Some people on YouTube have mounted the LIDAR Lite on a rotating platform but I have no idea if the data is high-quality enough
02:01 PM climbing_up: I have a software background, not an electronics background, so all these capacitors and resistors are scary
02:06 PM weyland|yutani: climbing_up, current and frequency just want to kill you nothing to be afraid :D
02:58 PM z64555: somewhat related, I'm messing with bezier curves and trajectories
02:59 PM z64555: also the capacitors aren't what's dangerous, the inductors are. :D
02:59 PM z64555: hm, unless you're talking about open air plate caps...
03:00 PM z64555: but those are rare
03:00 PM weyland|yutani: z64555, im talking about 600V 50Hz
03:01 PM z64555: at what ampherage
03:01 PM weyland|yutani: 800A
03:02 PM z64555: oh yeah, definitely
03:02 PM weyland|yutani: it was a pump motor
03:02 PM z64555: jeez, that's a beefy cap
03:02 PM weyland|yutani: a very large one
06:32 PM anonnumberanon: https://www.youtube.com/watch?time_continue=877&v=9pcQYuZ34Yw
06:33 PM anonnumberanon: machine terrorist
06:42 PM z64555: hm. bezier paths are actually pretty straight forward
06:43 PM z64555: can be composed of two points and two vectors
06:43 PM z64555: (cubic)
06:44 PM z64555: start point and vector, end point and vector
06:45 PM z64555: a velocity vector, v, starts with an initial value equal to the start vector, and is then lerp'd to the end vector
06:49 PM z64555: hm, wait, I'm missing something
07:22 PM z64555: yeah, definitely missed it
09:09 PM rue_mohr: i need a potocol for a tractor rov
09:10 PM rue_mohr: my last thought was to use an 8 bit number that somehow specified a low resolution vector
09:16 PM robotustra: are you building a tractor?
09:30 PM rue_mohr: no an rov
09:30 PM rue_mohr: a tractor drive rov
09:30 PM rue_mohr: do you know what tractor drive is?
09:31 PM rue_mohr: a tractor drive has two bidirectional drives placed a distance apart, bun that run parallel to each other
09:32 PM rue_mohr: each track rotates the robot (or whatever) around the other track by a radius
09:32 PM rue_mohr: kinda
09:32 PM rue_mohr: the turn radius is along a line that goes thru the centre point of the two drives
09:32 PM rue_mohr: it can be zero to infinity
09:33 PM rue_mohr: but the device cannot move sideways
09:33 PM rue_mohr: I was thinking of using a signed 7 bits to express the turning radius
09:33 PM rue_mohr: from -infinity to + infinity
09:34 PM rue_mohr: and then I suppose the other bit would be which direction it was moving along that radius
09:34 PM rue_mohr: so, maybe it should be just done in two signed 3 bit numbers
09:36 PM rue_mohr: my cnc machine dropped a nut today
09:36 PM rue_mohr: seems it managed to plough thrualmost 6mm of aluminum at once
09:37 PM rue_mohr: good ole unattended machinery
09:37 PM rue_mohr: destroyed the job
09:37 PM rue_mohr: quite thouroghly tho
09:38 PM rue_mohr: sliced a line right thru the middle of it, all the way thru the 1/4" plate
09:38 PM rue_mohr: one 0.4mm layer at a time (less that first cut)
09:38 PM rue_mohr: I need a larger motor for my next machine
09:38 PM rue_mohr: but yaknow, it just occured to me
09:39 PM rue_mohr: I dont need to worry about torque at high speeds cause high speed is a small bit
09:39 PM rue_mohr: so, its not that detremental to use a pulley
09:39 PM rue_mohr: cause big motors dont do high rpm
09:41 PM rue_mohr: typedef struct
09:44 PM rue_mohr: int leftDrive : 4;
09:44 PM rue_mohr: int rightDrive : 4;
09:44 PM rue_mohr: 0-100% duty
09:45 PM rue_mohr: but what do I use the two zeros for?
09:59 PM z64555: two zeros?
09:59 PM rue_mohr: -0 and +0
10:00 PM z64555: int's only have one zero. floats are the ones that have two
10:30 PM robotustra: 0 means stop?
10:31 PM anonnumberanon: z64555, what is the transfer function of a PID?
10:32 PM anonnumberanon: what do you do if someone asks you this during an interview
10:33 PM evilroot: I say I hate MatLab
10:33 PM anonnumberanon: lol
10:33 PM z64555: that's the sum of the various bits
10:33 PM z64555: uh, which domain we working in? probably easiest is in Laplace
10:34 PM evilroot: Yeah, see you've already failed
10:34 PM evilroot: Talk about derivative action and control values
10:34 PM z64555: Kp * F(s) + Ki * sF(s) + Kd * 1/s F(s)
10:34 PM ace4016: transfer function of a PID seems...
10:34 PM z64555: i think
10:34 PM ace4016: i mean, it's literally why it's called a PID
10:36 PM z64555: The PID is a transfer function, a specific one. I understood it as asking for what its equation is
10:36 PM ace4016: that's the transfer function. a proportional amplification + an integral gain + differential gain
10:36 PM ace4016: seems a bit odd of an interview question
10:36 PM ace4016: kind of like asking what's the equation of a quadratic formula
10:36 PM evilroot: Not if you're working in industrial controls
10:37 PM z64555: how so?
10:37 PM evilroot: For starters, what is the point of a PID?
10:38 PM ace4016: simple closed loop control that plebs understand :P
10:38 PM evilroot: NO
10:38 PM evilroot: It's far, far more than that
10:38 PM z64555: it is to control a machine or device to a desired output
10:38 PM * ace4016 waits for it
10:38 PM evilroot: And no at all simple
10:38 PM evilroot: z64555: by doing what?
10:39 PM z64555: It is part of a feedback control system where the machine's output is compared with its input (the desired target) to see how far away that machine is from reaching the target
10:40 PM evilroot: Again, its more than that
10:40 PM z64555: That error signal is then fed into the PID, which then given to the machine
10:40 PM evilroot: You're missing the I part
10:40 PM z64555: no I didn't? Ki * sF(s)
10:41 PM evilroot: What is the I factor?
10:41 PM z64555: integration.
10:41 PM z64555: This is in Laplace domain
10:41 PM evilroot: What does that actually mean? In terms of PID, what is I to the controller?
10:41 PM evilroot: Can you explain it in layman's?
10:42 PM jason__: I think that description only really covers the proportional part
10:42 PM evilroot: What are you "integrating"?
10:42 PM z64555: It is the continous-time accumulation of past error signals.
10:42 PM evilroot: Which could be called . . . .
10:43 PM z64555: "memory," if that's what you're trying to get at
10:43 PM evilroot: Exactly!
10:43 PM ace4016: ...
10:43 PM jason__: yeah I didn't see where you were goign with that
10:43 PM jason__: *going
10:43 PM ace4016: there are other control techniques of the adaptive/robust type that have all those properties
10:44 PM evilroot: PID . . . each aspect is how to best fix difference in where you want to be and where you are
10:44 PM ace4016: they're more involved however
10:44 PM evilroot: P is the current difference. AKA if its too cold, apply heat. Too hot, remove it
10:44 PM evilroot: And it's Proportional, aka if you're WAY too cold you turn the heater to max
10:44 PM jason__: there seems nothing special about the term "memory" compared to accumulated error to me
10:45 PM evilroot: I is past data. What the differences were in the past and what applying more or less resulted in. Did you overshoot? etc
10:45 PM evilroot: And D is anticipating future differences
10:46 PM evilroot: Given those three things you can hold a fairly steady value despite multiple external variables
10:46 PM z64555: yes, I think most of the channel knows this from looking at the equation
10:46 PM evilroot: Do they? Equations only really speak to some people
10:46 PM z64555: like engineers. Like the engineer that's interviewing you
10:47 PM evilroot: When I've interviewed people for jobs I've asked them to explain something to me as a normal person
10:47 PM ace4016: good idea
10:47 PM evilroot: I'm an engineer. I know the equations . . . but that's different from understanding the system
10:47 PM ace4016: though depending on how you do it, and about what, can be useless info >_>
10:47 PM evilroot: And how it works, what it really does
10:48 PM z64555: this is a good point, but is completely a tangent from the original question
10:48 PM evilroot: Not really.
10:48 PM evilroot: You asked about domains and gave an equation, right?
10:49 PM z64555: He simply asked "what is the transfer function of a PID," with no other information
10:50 PM evilroot: Granted that would be a stupid question
10:50 PM evilroot: Because who memorizes that :-D
10:50 PM z64555: folks that have to implement them on various archetectures
10:50 PM evilroot: Sorry for the tangent, interviews were something I spent a lot of time working on after getting really shitty applicant
10:50 PM evilroot: *applicants
10:51 PM z64555: I can tell. :)
10:51 PM z64555: It does seem like a real knee-jerker
10:51 PM z64555: "OMG HOW CAN YOU NOT KNOW THIS?"
10:52 PM z64555: or others that give the equation but not know what it means
10:52 PM evilroot: I owned a repair business
10:52 PM z64555: neat
10:53 PM evilroot: Its about more than just finding and fixing the problem. Explaining it to the customer in a way they can understand is what makes them trust you
10:53 PM evilroot: But really when it comes to any system, no matter how complex, on a basic level it is a collection of simpler machines
10:54 PM z64555: well yeah
10:54 PM evilroot: For instance with PID, do you know what application almost every single person owns?
10:54 PM z64555: I had a fun time the first time I was learning Kalman filters
10:54 PM evilroot: Cruise Control
10:55 PM z64555: Not just the cruise control
10:55 PM z64555: the engine throttle, too
10:55 PM evilroot: Yep
10:55 PM anonnumberanon: yeah the question was someone who applied at Tesla for software engineering, he didn't get it...
10:55 PM evilroot: Ahh, interesting
10:55 PM anonnumberanon: I was ready the Glassdoor job interview reviews today
10:55 PM evilroot: Not something the average code monkey would know
10:56 PM anonnumberanon: yeah it's an electrical engineering thing
10:56 PM evilroot: Seems a tad unfair really
10:56 PM anonnumberanon: i learned it a few years ago in college but don't remember how to use it
10:56 PM z64555: Yeah, most of the CompSci practice problems I come across are mathematics, graphs, an big data crunching
10:56 PM evilroot: I dual majored in comp sci and comp eng, but honestly mostly hated both
10:57 PM Tom_L: so why did you do it?
10:57 PM evilroot: Especially coding. I can do it, but it's like going to the dentist
10:57 PM anonnumberanon: z64555, yeah that sounds like the test i took today for a job
10:57 PM anonnumberanon: job interview
10:57 PM anonnumberanon: that was their screening process
10:57 PM evilroot: Tom_L: actually I quit, heh
10:57 PM anonnumberanon: I didn't do too well overall but got 100% on the first one.
10:57 PM evilroot: Started a business instead
10:58 PM anonnumberanon: overall 33% because I couldn't finish the other 2 on time
10:58 PM evilroot: And have been quite happy not being in a cubicle
10:58 PM anonnumberanon: trying to work them out anyway right now because they were interesting
10:58 PM evilroot: Currently I do a great deal of work with robotics . . . linear motion systems, CAM, etc
10:59 PM evilroot: I have surplus industrial robots I refurbished and converted building 3D printer components
10:59 PM z64555: oh that reminds me, there was a variation of PID's which used a different method for the D term
10:59 PM z64555: can't seem to remember the specifcs, though
11:00 PM evilroot: As well as various jobs from companies to consult
11:00 PM evilroot: But mostly I just love working on cool shit
11:02 PM anonnumberanon: yeah man, s'all good
11:03 PM anonnumberanon: i do need to make a massive move in programming
11:03 PM anonnumberanon: i have been practicing all day the past 4 days in preparation for this test
11:03 PM anonnumberanon: still didn't pass 50%, and i thought i was decent
11:03 PM anonnumberanon: im shit still
11:03 PM anonnumberanon: hopefully that will change
11:04 PM z64555: what kind of problems where you facing?
11:04 PM z64555: er, questions
11:04 PM z64555: I'm not too great at talking in multiple convo's at once
11:04 PM anonnumberanon: well the one that i got 100% it was about debugging already written code
11:04 PM anonnumberanon: was about a coin combination problem, heads or tails type of deal
11:05 PM z64555: experiance is the best teacher for debugging. :)
11:05 PM anonnumberanon: the 2nd one was some array thing representing teamperatures and you had to determine at which index summer started , i hated that one cause the way they presented it was crap
11:06 PM anonnumberanon: the last one though was awesome but I would have needed at least 2 hours extra to finish it
11:06 PM anonnumberanon: last was about finding all subsequences of an array and how to have 3 sums of any of its still available elements, and all sums must be equal
11:08 PM z64555: I don't understand that last one, would have to look at it
11:10 PM z64555: The temperatures thing might've been solved with a rolling average
11:10 PM z64555: probably not the best method, but it would allow for variations
11:11 PM anonnumberanon: it was really weird the way they stated it
11:11 PM anonnumberanon: ill tell you more hang on ill be back in 20 minutes when im on linux
11:12 PM * z64555 goes back to playing with beziers in the mean time