#robotics Logs

Apr 27 2020

02:55 AM deshipu: rue_mohr: I've seen a lot of algorithms that have been developed on a simulation first, and then applied to real robots
03:04 AM deshipu: sure, but still was much less work overall
03:05 AM rue_mohr: simulators are good for making things that work in simulators, be it circuits, or mechanics, or chemistry or robotics
03:06 AM deshipu: believe it or not, there is an overlap
03:07 AM rue_mohr: I understand that things that can only operate in simulators are important for the academic world so that students can get grades
03:07 AM rue_mohr: but if you want to make something that works in the real world, you need to make it in the real world, even if its just code.
03:08 AM deshipu: simulators are in the real world
03:09 AM rue_mohr: its just sad that people persist to state that simulated things have to work in the real world
03:10 AM deshipu: I wouldn't worry about people if I were you, they are not real
03:11 AM rue_mohr: I been watching for years, as people simulate things and then get confused when they dont really work
03:13 AM rue_mohr: I mean, I know this is all new to every person who encounters it, but they wont listen
03:14 AM rue_mohr: imagine watching a lineup of people, all slowly walking off to the right, and each one of them trips on a fold in the rug
03:14 AM rue_mohr: and you start telling each one of them to be carefull of the upcomming fold in the rug as its a tripping hazard
03:14 AM rue_mohr: and each one of them says that silly and then proceeds to trip on it
03:14 AM deshipu: yeah, so your solution is to sit there and loudly complain about how stupid humanity is
03:14 AM rue_mohr: thats my existence
03:15 AM rue_mohr: I'm just trying to tell you not to trip on the rug
03:15 AM rue_mohr: they wont let me fix it
03:15 AM deshipu: keep telling them how stupid they are, that should prevent them from tripping
03:15 AM rue_mohr: I'm not certified
03:16 AM rue_mohr: I'm so glad you totally understood what I said
03:16 AM deshipu: at least it makes you feel better about yourself
03:16 AM rue_mohr: so, just down the road there is going to be a snag with your simulation...
03:17 AM rue_mohr: no, nobody every takes and regard to anything I say
03:17 AM veverak: because you are wrong, it's that simple ;)
03:17 AM rue_mohr: most of the time they dont even hear me say it
03:18 AM rue_mohr: oh yea, and they always say I'm stupid and wrong
03:18 AM veverak: I never said stupid
03:18 AM veverak: (I know you are not)
03:19 AM rue_mohr: there is a fold in the carpet, fall on your face when you get there, its a part of your life you have made unavoidable
03:20 AM rue_mohr: http://users.tpg.com.au/users/ldbutler/NegResDipMeter.htm
03:20 AM veverak: you do not need real robot to test and verify that your protocol serialization works, and doing that in 'tests' is less painfull than in the hardware itself
03:20 AM rue_mohr: negitive resistance, thats cool
03:21 AM veverak: got lua virtual machine that works as interpreter of commands, debuggin half of it in simulation saved time, and yet I did not have to throw it out in the real robot, just debug different bugs :P :)
03:21 AM rue_mohr: veverak, when it dones't work as it did when you simulated it, keep is quiet so you dont prove me right, not that anyone will remember I said it
03:23 AM veverak: no, I won't, I can admit when I fail
03:23 AM veverak: I used simulation and it is handy tool to speedup dvelopment if used correctly
03:23 AM veverak: but I don't think you are willing to change your world-view, so meh
03:26 AM veverak: which is exactly the problem here, the key to understanding is stop thinking about "does simulation work or not?" and rather ask "when, under which condition, and for what it works?"
03:27 AM veverak: somehow you are stuck in the "does it work or not?" kind of thinking and that simply won't do it
03:29 AM rue_mohr: I learned to be mindfull of simulators when using circuit simulators, and as I watched EVERYONE try to use all sorts of simulators, and have their projects fail when they hit the real world, I developed my opinion. Save time, make it in the real world to begin with.
03:31 AM veverak: Yes, I am saving time with my experience of this stuff - where simulation is beneficial
03:31 AM veverak: and if you think that all simulation is bad because circuit-simulators failed, well than THAT is the problem
04:13 AM Jak_o_Shadows: Needs to be a quick feedback loop between simulation and hardware
04:13 AM Jak_o_Shadows: No point delaying that too long, that's when problems happen.
04:13 AM * Jak_o_Shadows is fighting that fight at work
04:59 AM deshipu: sure, simulation and hardware should be informed by each other
04:59 AM deshipu: but jumping right into hardware, a la James Bruton, is going to make you waste a lot of time and money
05:57 AM veverak: Jak_o_Shadows: that fist my experience :)
05:58 AM veverak: *fits
06:00 AM Jak_o_Shadows: Yeah. We pay a bit too much lip service to it at work
06:00 AM Jak_o_Shadows: deshipu, james bruton frustrates me
06:01 AM deshipu: Jak_o_Shadows: same
06:16 AM Jak_o_Shadows: If I had half as much time.
07:13 AM LarchOye1 is now known as Guest18995
09:27 AM rue_mohr: veverak, you missed the part where I mentioned that over the 18 years I been on irc, everyone who has tried to simulate something has told stories of how it didn't work out
11:35 AM -!- #robotics mode set to +o by orwell.freenode.net
11:46 AM -!- #robotics mode set to +o by orwell.freenode.net
12:30 PM -!- #robotics mode set to +o by orwell.freenode.net
12:37 PM srk: sup nerds
12:42 PM -!- #robotics mode set to +o by orwell.freenode.net
12:43 PM LarchOye1 is now known as Guest46986
03:47 PM LarchOye1 is now known as Guest27644
04:17 PM LarchOye1 is now known as Guest17700
08:19 PM zhanx_ is now known as zhanx