[00:11] <m1x10> gnite all
[00:11] m1x10 (m1x10@ppp089210140086.dsl.hol.gr) left irc: Quit: I eat electrons for breakfast and I vomit thunders
[00:13] natrium42 (~natrium@CPE000625d867e2-CM0014045885be.cpe.net.cable.rogers.com) joined #highaltitude.
[00:30] SpeedEvil (~user@tor/regular/SpeedEvil) left irc: Ping timeout: 260 seconds
[00:31] <natrium42> o/
[00:32] <fsphil> morning natrium42 :)
[00:32] <natrium42> evening fsphil :)
[00:32] <fsphil> haha
[00:41] dcrand (~d_crand@68-117-98-7.dhcp.eucl.wi.charter.com) joined #highaltitude.
[00:41] dcrand (d_crand@68-117-98-7.dhcp.eucl.wi.charter.com) left #highaltitude.
[00:42] <fsphil> night all
[00:42] fsphil (~phil@beastie.sanslogic.co.uk) left irc: Quit: Spooooock
[00:50] SpeedEvil (~user@tor/regular/SpeedEvil) joined #highaltitude.
[00:58] hyte (519f817a@gateway/web/freenode/ip. joined #highaltitude.
[01:08] hyte (519f817a@gateway/web/freenode/ip. left irc: Quit: Page closed
[01:15] SpeedEvil (~user@tor/regular/SpeedEvil) left irc: Ping timeout: 245 seconds
[01:39] MoALTz (~no@ joined #highaltitude.
[01:45] SpeedEvil (~user@tor/regular/SpeedEvil) joined #highaltitude.
[01:47] <Wild-Wing> hi
[01:54] jasonb (~jasonb@dsl092-009-225.sfo1.dsl.speakeasy.net) left irc: Ping timeout: 265 seconds
[01:56] natrium42 (~natrium@CPE000625d867e2-CM0014045885be.cpe.net.cable.rogers.com) left irc: Quit: Leaving
[02:20] juxta_ (blah@ppp59-167-4-96.lns1.mel4.internode.on.net) joined #highaltitude.
[02:26] Wild-Wing (~chatzilla@c-24-34-60-55.hsd1.ma.comcast.net) left irc: Remote host closed the connection
[02:33] juxta_ (blah@ppp59-167-4-96.lns1.mel4.internode.on.net) left irc: Ping timeout: 276 seconds
[02:47] MoALTz (~no@ left irc: Quit: Leaving
[03:22] natrium42 (~natrium@CPE000625d867e2-CM0014045885be.cpe.net.cable.rogers.com) joined #highaltitude.
[05:02] juxta_ (~blah@ joined #highaltitude.
[05:11] juxta_ (~blah@ left irc: Ping timeout: 240 seconds
[05:40] jasonb (~jasonb@m480536d0.tmodns.net) joined #highaltitude.
[06:08] jasonb_ (~jasonb@m390536d0.tmodns.net) joined #highaltitude.
[06:11] jasonb (~jasonb@m480536d0.tmodns.net) left irc: Ping timeout: 246 seconds
[06:18] jasonb_ (~jasonb@m390536d0.tmodns.net) left irc: Ping timeout: 276 seconds
[06:38] Upu (~Upu@ubn.upuaut.net) joined #highaltitude.
[06:48] natrium42 (~natrium@CPE000625d867e2-CM0014045885be.cpe.net.cable.rogers.com) left irc: Quit: Leaving
[07:29] fsphil (~phil@beastie.sanslogic.co.uk) joined #highaltitude.
[07:49] gamers600 (~gamers600@ip68-99-230-95.ph.ph.cox.net) joined #highaltitude.
[07:53] Simon-MPFH (~simon@phantom.mpfh.co.uk) joined #highaltitude.
[08:17] jasonb_ (~jasonb@adsl-66-124-73-250.dsl.sntc01.pacbell.net) joined #highaltitude.
[08:49] gamers600 (~gamers600@ip68-99-230-95.ph.ph.cox.net) left irc: Quit: gamers600
[08:55] m1x10 (m1x10@ppp089210140086.dsl.hol.gr) joined #highaltitude.
[08:55] <m1x10> morning
[08:58] <m1x10> ping LA3QMA
[08:59] <LA3QMA> m1x10: i dont have much time now. but i think thers an error in your code or that you are sending the same static information to often and that is why you dont show up in aprs.fi all the time
[08:59] <m1x10> LA3QMA
[08:59] <m1x10> Its fixed now
[09:00] <m1x10> SW2HYX-11>APZIKA,WIDE2-1,qAR,SZ7SER:>40.6533/22.9438/000415/172/002/+000/06/4.6/+31/43/0733/0639/IKAROS
[09:00] <LA3QMA> anyhow strange that the digipeater closer to you is not responding
[09:00] <LA3QMA> can you just for fun of it test with TRACE2-1 ??
[09:00] <m1x10> ok
[09:00] <m1x10> wait
[09:02] Simon-MPFH (~simon@phantom.mpfh.co.uk) left irc: Quit: Leaving
[09:04] <m1x10> SW2HYX-11>APZIKA,TRACE2-1,qAR,SZ7SER:>40.6481/22.9444/000209/206/005/+000/07/4.5/+32/40/0000/0802/IKAROS
[09:09] <LA3QMA> strange i would thing the closer digipeater should have done something
[09:09] <LA3QMA> i have to go now. i'm back later . anyhow the package is correct
[09:10] <m1x10> :)
[09:10] <m1x10> ok see u
[09:29] <m1x10> ping fsphil, jonsowman, all :)
[09:29] <jonsowman> hello m1x10
[09:29] <m1x10> hello
[09:29] <fsphil> morning
[09:30] <m1x10> may I ask something about electronics?
[09:30] <m1x10> hello fsphil
[09:33] <m1x10> busy all ! Everybody works except me :(
[09:35] <juxta> m1x10, go ahead :)
[09:35] <juxta> don't ask to ask, just ask ;)
[09:35] <m1x10> hey juxta
[09:35] <m1x10> before I ask I have something for you
[09:36] Action: fsphil has been 'working' in VB all morning .. could use the brain exercise :)
[09:36] <m1x10> remember we wished to find a NMEA sentence in ublox that contains all the necessary info?
[09:37] <m1x10> well, if you still looking I got the solution
[09:37] <juxta> oh?
[09:37] <m1x10> :)
[09:37] <m1x10> ready?
[09:37] <juxta> sure
[09:37] <m1x10> u disable all NMEA once and for all
[09:37] <m1x10> uaGPS.println(GGA_OFF);
[09:37] <m1x10> uaGPS.println(GLL_OFF);
[09:37] <m1x10> uaGPS.println(GSA_OFF);
[09:37] <m1x10> uaGPS.println(GSV_OFF);
[09:37] <m1x10> uaGPS.println(RMC_OFF);
[09:38] <m1x10> and
[09:38] <m1x10> uaGPS.println("$PUBX,00*33"); !!
[09:38] <juxta> does that switch on UBX output?
[09:39] <m1x10> its just prints a sentence in simple ASCII
[09:39] <m1x10> containing everything !
[09:39] <fsphil> what does it print?
[09:39] <m1x10> PUBX,00,235952.00,0000.00000,N,00000.00000,E,0.000,NF,6496153,4593474,0.000,0.00,0.000,,99.99,99.99,99.99,0,0,0
[09:39] <m1x10> its not locked
[09:40] <m1x10> i used it the last day and its verrrrry cool
[09:40] <m1x10> gps will send you a response only if you request it
[09:40] <m1x10> thus not flooding the buffers
[09:41] <m1x10> so you void loop { uaGPS.println("$PUBX,00*33"); }
[09:41] <m1x10> its a supreme solution !
[09:41] <m1x10> I should edit some wiki
[09:42] <juxta> nice
[09:42] <juxta> are the fields documented somewhere?
[09:42] <m1x10> in brief contains: lat/lot/alt course/speed/verticalspeed satellite number
[09:42] <fsphil> very nice. how long do you wait for a reply?
[09:42] <m1x10> it comes instantly
[09:42] <m1x10> no sleeps or delays
[09:43] <m1x10> juxta I got the pdf
[09:43] <m1x10> wait
[09:43] <juxta> the UBX protocol spec?
[09:44] <m1x10> yes
[09:44] <m1x10> page 52
[09:44] <m1x10> Proprietary Messages
[09:44] <m1x10> Proprietary Messages : i.e. Messages defined by u-blox.
[09:44] <m1x10> UBX,00
[09:44] <juxta> ah yep I see it
[09:44] <juxta> I totally glossed over that, hehe
[09:44] <m1x10> the ublox port should be enabled to support both nmean/ublox
[09:44] <m1x10> but this is the default setting
[09:46] <m1x10> was my info useful ?
[09:47] <juxta> yep! :)
[09:47] Simon-MPFH (~simon@phantom.mpfh.co.uk) joined #highaltitude.
[09:47] <m1x10> the vertical speed is very nasty
[09:47] <m1x10> if you fall it becomes negative number !!
[09:47] <m1x10> if u go up positive
[09:48] <juxta> what's wrong with that?
[09:48] <m1x10> thats how it works
[09:48] <m1x10> :)
[09:48] <m1x10> you fall at 5m/s its shows -5 :)
[09:48] <juxta> that sounds OK to me, that's what I would expect
[09:49] <m1x10> you go up at 5m/s it shows 5 :)
[09:49] <m1x10> this way we dont need accelometers or stuff
[09:49] <m1x10> juxta now I forgot what I wanted to ask !
[09:50] <juxta> haha
[09:50] <m1x10> anyway is there any public wiki to paste this info
[09:50] <m1x10> ?
[09:50] <juxta> yes
[09:50] <juxta> ukhas.org.uk
[09:50] <juxta> look at the FSA03 page
[09:50] <m1x10> ok
[09:51] Jasperw1 (~jasperw@212-98-44-230.static.adslpremium.ch) left irc: Ping timeout: 245 seconds
[09:51] <m1x10> "mixio contributed at last to the community"
[09:51] <m1x10> :)
[09:53] <m1x10> How I login? I dont have account at that page
[09:53] <m1x10> can someone make account ?!
[09:53] <fsphil> click on the login box
[09:53] <fsphil> there should be a link for registering
[09:55] <m1x10> ok
[09:56] <m1x10> oh
[09:56] <m1x10> someones writes
[09:56] <m1x10> he antenna on this module appears to be slightly wobbly and this may be one of the causes of the module not getting a lock for long periods of time. Its recommended that you secure it (perhaps with glue) in a position that works.
[09:56] <m1x10> lol
[09:59] <m1x10> I have to press Recover draft in order to edit it?
[10:00] <fsphil> just the little edit button below the section you want to edit
[10:00] <fsphil> always preview your changes first :)
[10:05] <m1x10> also figure out that ZDA sentence is disabled on fsa03 by default
[10:11] eroomde (~ed@nessie.habhub.org) joined #highaltitude.
[10:11] <eroomde> jonsowman: and on a key now - don't need a passwd anymore
[10:11] <eroomde> sorry for the delay, just wanted a decent way to manage it acorss all the various machines
[10:13] Jasperw (~jasperw@212-98-44-230.static.adslpremium.ch) joined #highaltitude.
[10:24] <m1x10> everybody http://ukhas.org.uk/guides:falcom_fsa03 !!
[10:25] <eroomde> excellent contribution m1x10 !
[10:26] <m1x10> :)
[10:26] <eroomde> that's a really helpful page
[10:26] <juxta> hey, there may be other contibutors ;p
[10:26] <eroomde> ah :)
[10:27] <m1x10> haha
[10:27] <m1x10> juxta
[10:27] <m1x10> i remembered my question
[10:27] <eroomde> yes i just had a look at the history
[10:27] <eroomde> :o)
[10:28] <juxta> sure m1x10 - what was up?
[10:28] <juxta> eroomde is here now and he's pretty much the man so I think you'll get a good answer from someone at least :)
[10:28] <m1x10> fsa03 tx/rx is at 3.3v ok. my power source is 5v. So I use the potential divider circuit to send to rx safely.
[10:29] <juxta> yep
[10:29] <m1x10> but measuring the tx it shows ~5v
[10:29] <juxta> from the module?
[10:29] <m1x10> no
[10:30] <juxta> after the divider you mean?
[10:30] <eroomde> can someone just try sending me a message with my name? i want to see if i can get putty to flash the task bar in xp
[10:30] <juxta> eroomde, eroomde, eroomde, eroomde, eroomde
[10:30] <m1x10> no juxta
[10:30] <eroomde> woooooooo
[10:30] <eroomde> thanks juxta
[10:30] <eroomde> it works
[10:30] <m1x10> let me explain
[10:30] <juxta> haha
[10:30] <juxta> there's some bug with this version of xchat I am using
[10:30] <eroomde> ok cool, now I have interrupts rather than polling. this is progress
[10:31] <juxta> it keeps displaying a popup balloon telling me I have been highlighted, which was true about 4 hours ago
[10:31] <eroomde> better late than never I guess
[10:32] <eroomde> i just read your message that I am 'pretty much the man'. you make me blush, sir
[10:32] <m1x10> to safely send 5v data to 3.3v rx port of fsa03 i use the divider and works.
[10:32] <m1x10> fsa03 TX sends at 3.3v
[10:32] <m1x10> but that line ends to a 5v
[10:33] <eroomde> ok last thing I promise - could someone send me a pm? just to check that bit works too
[10:33] <m1x10> i measured the volts on that line and shows ~5v
[10:34] <m1x10> which is correct
[10:34] <m1x10> my question is
[10:34] <m1x10> when fsa03 TX will send something to this line
[10:34] <m1x10> what will happen to the volts of this line?
[10:35] <m1x10> bah stupid explanation
[10:35] <m1x10> suicide now.
[10:35] <juxta> m1x10, i understand what you man
[10:35] <juxta> mean*
[10:36] <juxta> that line will only go to 3v3 when the FSA03 is transmitting, but that's no problem - your 5V arduino will recognise 3v3 as a 'high' anyway
[10:36] <m1x10> yes
[10:36] <m1x10> are you sure that when the fsa will transmit the 5v line will go 3.3v?
[10:37] <juxta> yes
[10:37] <m1x10> ok
[10:37] <m1x10> now
[10:37] <m1x10> how u explain that
[10:37] <m1x10> electrically
[10:37] <m1x10> whats happening
[10:37] <juxta> the RX line on the arduino will be at 0V potential, not at 5V potential
[10:38] <m1x10> i measure 5volts. if fsa transmits it falls to 3.3v. Just like that?
[10:38] <juxta> current will flow from the TX pin on the FSA03 at a potential of 3v3 to the RX pin on the Arduino
[10:38] <m1x10> aaa wait
[10:39] <m1x10> I may measure 5v but there is no current
[10:39] <juxta> i would have thought it would be low potential. I havent measured though, it might be set high with high impedence, I dont know - eroomde, care to offer input? ;p
[10:40] <m1x10> and when fsa03 puts some current flowing the 5v fall to 3.3v
[10:40] <juxta> i've go to have a shower and get ready to head off - will catch you tomorrow m1x10 :)
[10:40] <m1x10> ok juxta
[10:40] <m1x10> you go to work?
[10:40] eroomde (~ed@nessie.habhub.org) left irc: Quit: leaving
[10:41] <juxta> haha, thankfully not! :)
[10:41] <m1x10> holiays ?
[10:41] <m1x10> :p
[10:41] <juxta> it's 7pm
[10:41] <m1x10> lol
[10:41] <m1x10> anyway
[10:41] <m1x10> cu !
[10:41] <juxta> :)
[10:47] eroomde (~ed@nessie.habhub.org) joined #highaltitude.
[11:01] <fsphil> ah, the nessie lives
[11:01] <eroomde> aye she does
[11:04] SpeedEvil (~user@tor/regular/SpeedEvil) left irc: Quit: Leaving.
[11:04] SpeedEvil (1000@tor/regular/SpeedEvil) joined #highaltitude.
[11:04] <jonsowman> cheers eroomde
[11:04] <jonsowman> will disable pwd auth
[11:05] <DanielRichman> too slow.
[11:05] <jonsowman> D:
[11:05] <DanielRichman> jonsowman: http://nessie.habhub.org/
[11:07] <jonsowman> nice work DanielRichman
[11:07] <DanielRichman> You haven't seen how it's generated yet. It's properly disgusting
[11:07] <DanielRichman> /etc/web-status-page
[11:08] <jonsowman> oh good lord
[11:08] <jonsowman> :P
[11:08] <DanielRichman> exactly
[11:10] juxta (juxta@ppp203-122-193-94.static.internode.on.net) left irc: Ping timeout: 240 seconds
[11:11] <m1x10> DanielRichman nice
[11:14] <fsphil> heh I was going to say I liked the nice dark brown background, but I just looked at the html and it's blue. my monitor sucks
[11:14] <fsphil> or my eyes, one of the two :)
[11:14] <DanielRichman> :o
[11:21] <m1x10> lol
[11:33] MoALTz (~no@ joined #highaltitude.
[12:27] <SpeedEvil> Sigh.
[12:27] <SpeedEvil> http://science.slashdot.org/story/10/08/05/2141255/SpaceX-Unveils-Heavy-Lift-Rocket-Designs
[12:27] <SpeedEvil> Makes me want to move to the USA, simply so I can yell really loudly at my congressperson.
[12:29] <Randomskk> jonsowman: eroomde: around?
[12:30] <Randomskk> DanielRichman: might be worth having IRC nicks on that page as well as names
[12:30] <DanielRichman> OK
[12:30] <Randomskk> perhaps a team->root alias as well or something similar
[12:30] <Randomskk> the page does look excellent though I have not yet seen how you are generating it :P
[12:31] <DanielRichman> I'll pastebin the generator if you feel like throwing up :P
[12:31] <Randomskk> yay go for it
[12:31] <DanielRichman> btw Randomskk the ntp server configured in ntp.conf is a tier 1 box in the colo (GPS)
[12:31] <DanielRichman> so do we need the UK pool?
[12:32] <Randomskk> ah, no, not really then
[12:32] <Randomskk> well
[12:32] <Randomskk> does it hurt though?
[12:32] <Randomskk> backup etc
[12:33] <DanielRichman> ntpd should work out that the t1 server is much higher quality and provided it responds, ignore all the others?
[12:33] <DanielRichman> http://pastie.org/1079377
[12:33] <Randomskk> I believe so
[12:34] <Randomskk> and thus if it doesn't respond it'l use pool
[12:34] <Randomskk> worth having multiple servers configured at any rate
[12:34] <Randomskk> oh good lord
[12:34] <russss> apparently you should have at least 3 servers in the NTP config
[12:34] <DanielRichman> yup. At the start it seemed like a good idea
[12:34] <DanielRichman> now... not so much
[12:34] <Randomskk> I think pool counts for enough though you could 0.uk.pool.ntp.org etc
[12:35] <DanielRichman> Randomskk: it's still more efficient than php
[12:35] <Randomskk> eeeh
[12:35] <Randomskk> yes maybe
[12:35] <Randomskk> cronjob?
[12:35] <DanielRichman> yeah at the bottom
[12:35] <Randomskk> yes, I see
[12:35] <russss> Randomskk: if you only have 2 servers and one goes wrong which one can you trust? :)
[12:35] <Randomskk> well, lol
[12:35] <DanielRichman> :D
[12:35] <Randomskk> russss: true
[12:36] <DanielRichman> well if the t1 one goes wrong then ntp will blindly follow it, no?
[12:36] <Randomskk> I think it's basically very smart and smart enough to handle being told lots of servers
[12:36] <russss> DanielRichman: nah it's quite clever
[12:36] <DanielRichman> How about we make ntp.ubuntu.com the third server?
[12:36] <Randomskk> go for it
[12:36] <russss> just use 0.europe.pool.ntp.org and 1.europe.pool.ntp.org
[12:37] <Randomskk> why not uk.pool
[12:37] <Randomskk> I think ntp.ubuntu is just an alias for ntp.org?
[12:37] <russss> Randomskk: I didn't know that existed :)
[12:37] <Randomskk> no, it's not
[12:37] <Randomskk> russss: it does
[12:37] <Randomskk> :P
[12:37] <Randomskk> 0.uk.pool.ntp.org. 270 IN A
[12:37] <Randomskk> 0.uk.pool.ntp.org. 270 IN A
[12:37] <Randomskk> 0.uk.pool.ntp.org. 270 IN A
[12:37] <russss> Randomskk: but my gut feeling is that you get more variety with european servers without sacrificing too much latency.
[12:37] <Randomskk> true I guess
[12:37] <russss> it's a bit of an academic discussion though :P
[12:38] <Randomskk> it is rather
[12:38] <Randomskk> the tier 1 in the datacentre will probably never go wrong anywya
[12:38] <DanielRichman> daniel@nessie:~$ ntpq -p
[12:38] <DanielRichman> http://pastebin.com/kZybGxVU
[12:38] <Randomskk> is that good? bad? okay?
[12:38] <russss> after 10-20mins a star will appear next to one of those servers, and that's the one which it's using as primary
[12:39] <DanielRichman> cool
[12:39] <DanielRichman> Aug 6 12:37:01 nessie ntpd[28545]: frequency initialized -32.946 PPM from /var/lib/ntp/ntp.drift
[12:39] <russss> ubuntu actually gives you a choice of servers now, and I don't think pool is among them
[12:40] <russss> I selected three and it seems like they're all tier one, heh
[12:40] <russss> http://pastebin.com/txWz3wHm
[12:40] <DanielRichman> heh
[12:45] <jonsowman> am vaguely around russss
[12:46] <jonsowman> * Randomskk
[12:46] <jonsowman> sup?
[12:49] <DanielRichman> the gps source now has a *
[12:49] <DanielRichman> :)
[12:49] <Randomskk> jonsowman: oh sorry hi
[12:49] <Randomskk> the email
[12:50] <Randomskk> cusf
[12:53] <DanielRichman> Any idea why auth.log has Just Stopped?
[13:04] Action: DanielRichman restarts rsyslogd
[13:04] <DanielRichman> hmm; is back now
[13:08] MoALTz (~no@ left irc: Read error: Connection reset by peer
[13:08] MoALTz (~no@ joined #highaltitude.
[13:10] Simon-MPFH (~simon@phantom.mpfh.co.uk) left irc: Quit: Leaving
[13:27] <jonsowman> Randomskk: hmm yeh
[13:27] <jonsowman> cant even remember what date im coming back to cam
[13:27] <jonsowman> i might be the weekend after you
[13:32] <Randomskk> I am the 'week early' option as it were
[13:32] <Randomskk> come up 25th
[13:32] <Randomskk> term starts 5
[13:32] <Randomskk> so I have the weekend, the following week and weekend, then we start on tuesday
[13:32] <Randomskk> if you're coming up the week after you have the weekend and money and then term starts tuesday
[13:32] <Randomskk> lectures start thursday I guess
[13:32] <Randomskk> 7
[13:33] <Randomskk> at a guess
[13:33] <Randomskk> dunno if ed's in aus though
[13:35] <Randomskk> jonsowman: ping
[13:35] LA3QMA (kgb@57.62-97-229.bkkb.no) left #highaltitude.
[13:45] <jonsowman> Randomskk: pong
[13:45] <Randomskk> ^
[13:45] <jonsowman> cool okay
[13:45] <jonsowman> i think im coming up 2nd
[13:45] <jonsowman> duno atm
[13:46] <Randomskk> okay
[13:48] <jonsowman> bbl
[13:48] <m1x10> is there any tricky way to add some more RAM to my aprs system?
[13:49] <m1x10> atmega328's ram is 2kb and its over !
[13:54] <eroomde> m1x10: you'll have to refactor
[13:54] <m1x10> what do u mean by that?
[13:54] <Randomskk> make your code use less ram :P
[13:54] <eroomde> well, not stricly true, you can write things to flash or to some external memory, but thatr's a big big faff. you should be able to fit a system such as this into the ram that you have available, with some careful analysis of your code
[13:55] <m1x10> yes I have done too much analysis
[13:55] <m1x10> get go lower
[13:55] <m1x10> i just added the pressure sensor in the aprs
[13:55] <Randomskk> 2k is still plenty
[13:55] <Randomskk> there's no way to add more ram. you can add external flash memory but that may not really help
[13:55] <m1x10> and with its code i run of
[13:56] <Randomskk> need a beefier micro for more ram, but as eroomde said you should be able to fit this system into that much ram
[13:56] <eroomde> yes, that's >2000 bytes of variables you've got floating about - that's really quite a lot!
[13:56] <Randomskk> are you running out of ram, or out of code flash space?
[13:56] <m1x10> RAM
[13:56] <Randomskk> eroomde: any thoughts on the email from that eads guy?
[13:56] <Randomskk> m1x10: how do you know?
[13:56] <m1x10> I count it
[13:56] <eroomde> Randomskk: yes, i did email him back at the time but it seems it's blocked
[13:57] <m1x10> without pressure sensor code free ram get 54byte
[13:57] <eroomde> my thoughts: we provide launch site and consultancy on what to do 9with an emphasis of safe and light payloads) and they provide everything else, including a donation to cusf
[13:57] <Randomskk> sounds good
[13:57] <m1x10> anyway
[13:57] <m1x10> must go
[13:57] <Randomskk> however email being blocked isn't great
[13:57] m1x10 (m1x10@ppp089210140086.dsl.hol.gr) left irc: Quit: I eat electrons for breakfast and I vomit thunders
[13:57] <Randomskk> I enjoyed his email recall as well
[13:58] <eroomde> lol
[13:58] <Randomskk> I didn't know email worked like that
[13:58] <eroomde> better than running into the server room and hitting the emergency power cut button
[13:58] <eroomde> as someone at my work once did when they sent an embarassing email to the wrong person
[13:58] <Randomskk> haha nice
[13:58] <eroomde> sadly they couldn't outrun email
[13:58] <Randomskk> did they get it in time?
[13:58] <Randomskk> awww
[13:58] <Randomskk> gmail has 'Undo' now which I also enjoy
[14:00] <DanielRichman> So I was thinking about the consolidation of all the various tools
[14:01] <DanielRichman> and since all of them use a map, I was thinking that we could organise it such that the "homepage" is a static file, which includes several JS scripts including a mapping abstraction layer to swap between gmaps & OS, and a JS script for each tool. And each tool's script knows where it will be AJAXing to
[14:01] <DanielRichman> so the tracker script will contact couchdb; the predictor will contact the python/php glue in the predict folder
[14:01] <eroomde> that sounds sensible
[14:01] <DanielRichman> And then each "tool" creates a rounded-corner floaty box, sortof the same style as the predictor v2 is
[14:02] <DanielRichman> that way useful things such as having multiple predictions, or multiple predictions, a track, and a flight archive, plotted on one map can happen
[14:03] <DanielRichman> Also, for the tracker, I think we should have a database of flights; where we specify the time range and the callsign used. Anything that doesn't match that goes into a sandbox/testing "track" so that people testing out dl-fldigi don't spam up the homepage
[14:03] <DanielRichman> then you select which tracks you want to view.
[14:03] <DanielRichman> Also when the timerange is up that flight goes into the archive.
[14:03] <DanielRichman> That also means that a "next-flight" box can be automagic
[14:10] <fsphil> is it wise having so much on the front page?
[14:11] <DanielRichman> minimisable boxes?
[14:13] <Randomskk> a menu
[14:14] <Randomskk> 'open predictor' etc
[14:14] <Randomskk> default to showing tracks I guess
[14:14] <Randomskk> I like this plan
[14:14] <fsphil> yea
[14:14] <DanielRichman> it could default depending on what's going on
[14:14] <DanielRichman> ie if there's a flight planned for today then yeah, default to tracker
[14:14] <DanielRichman> otherwise the archive or the predictor maybe
[14:14] <DanielRichman> perhaps user customizable
[14:14] <fsphil> even better
[14:14] <Randomskk> well what do we have: achive, current data, predictor?
[14:15] <Randomskk> archive* even
[14:15] <DanielRichman> past present future
[14:15] <Randomskk> hah, quite
[14:15] <Randomskk> so, default to present if there is any, if not default to the last archived flight
[14:15] <Randomskk> archive box selects callsign and flight
[14:15] <Randomskk> we want some way of grouping records into a 'flight' rather than a timerange I think
[14:16] <Randomskk> predictor box runs a prediction, and current data just shows everything in the current flight - probably its box would have a callsign dropdown in the event of multiple flights, defaulting to 'all' perhaps
[14:16] <DanielRichman> you could have an archive box, you select then click "show on map"... can show multiple traces on the map at once
[14:16] <Randomskk> yea
[14:16] <Randomskk> well
[14:16] <Randomskk> maybe
[14:16] <DanielRichman> ie see where icarus 1, 2, 3, 4 went etc.
[14:16] <Randomskk> second thought
[14:16] <Randomskk> a 'tracks' box on the screen that lists all current traces on the map and you can show/hide/remove any
[14:17] <Randomskk> predictor adds a Prediction track, archive adds an Archive track, the current data is its own track, one per callsign
[14:17] <DanielRichman> yup
[14:17] <DanielRichman> we also need to find a way to include the hourly-predictor feature
[14:17] <Randomskk> so the archive box lets you see old tracks, the predictor box lets you generate new tracks
[14:17] <Randomskk> yea
[14:17] <Randomskk> that can be an option on the predictor box I guess, show tracks for next x hours, all coming up as one entry on the track list
[14:18] <DanielRichman> it's sort of like a track but not, that's the problem
[14:18] <eroomde> bear in mind there's a reasonable amount of computation and downloading for a 120hr hourly-prediction
[14:18] <Randomskk> you specify your launch parameters as before, but tell it to generate N hours starting at date M
[14:18] <DanielRichman> eroomde: prediction caching is a must
[14:18] <Randomskk> DanielRichman: it's a group of tracks then, support nesting on the track list
[14:18] <Randomskk> eroomde: computation is minimal
[14:18] <Randomskk> data is more significant
[14:18] <DanielRichman> Randomskk: and then have the option to only display the point where it landed
[14:19] <Randomskk> eroomde: the hourly predictions take about 20 seconds to run if all data is cached
[14:19] <eroomde> that's non-trivial
[14:19] <Randomskk> DanielRichman: sounds good
[14:19] <Randomskk> eroomde: true
[14:19] <eroomde> although mnaybe on a xeon
[14:19] <eroomde> :)
[14:19] <Randomskk> well, there's only so much we can do about that anyway
[14:19] <Randomskk> like only authorised users, or at most 3 days worth, or decaying the interval
[14:19] <DanielRichman> eroomde: BTW; on a side note; can I suggest the use of aptitude rather than apt-get on nessie so that we have the automatically-installed flags to mark stuff explicitly/notexplicitly installed; just to keep things tidier
[14:19] <Randomskk> so hourly for 12 hours, bi-hourly for the next 12, every 4 for the next 24, etc
[14:20] <eroomde> yep i've already been told about that
[14:20] <eroomde> will do that from now on
[14:20] <Randomskk> I think hourly decaying would probably look better anyway
[14:20] <eroomde> well, as someone who uses it a lot, I actually like all 120hrs straight up
[14:21] <Randomskk> fair enough. I guess either have an option for it or make it for authorised users only
[14:21] <DanielRichman> How many people are going to be realistically hitting the hourly predictor?
[14:21] <Randomskk> we can't usefully cache it given its variable parameters, and we can't reduce the computation time if we want 120 predictions
[14:21] <DanielRichman> If it gets slashdotted then I guess lots of people are gonig to press the button out of curiosity
[14:21] <Randomskk> I'd say the default could happily be less, but besides that
[14:21] <SpeedEvil> It could get slashdotted during a flight too
[14:22] <eroomde> yes, that's probably the major use case
[14:22] <DanielRichman> besides if it got slashdotted you'd have people spamming the standard predictor anyway
[14:22] <DanielRichman> which would have the same, bad, effect
[14:22] <eroomde> i don't think people are that interested in predictors
[14:22] <eroomde> it''s not like the predictor would get slashdotted
[14:22] <DanielRichman> well true, but flights can be quite slow sometimes and if there's a shiny predict button...
[14:22] <DanielRichman> I guess if that happens we'd just have to disable it.
[14:23] <eroomde> that's a fail
[14:23] <DanielRichman> ino
[14:23] <Randomskk> I'm sure we could manage if it came to it
[14:23] <Randomskk> disable predictions for unauthorised users when 1min load is above 2 or whatever
[14:23] <DanielRichman> There are various features of PAM/linux to put a cap on the amount of ram/number of processes/amount of cpu used by one user
[14:24] <eroomde> but yeah, perhaps just have 120hr cron as with now and then some other less computationally expensive, more user-friendly thing
[14:24] <Randomskk> worth having some form of netbook mode that makes everything small and minimises as much as possible too
[14:24] <Randomskk> eroomde: 120hr cron has the disadvantage of being fixed parameters
[14:24] <DanielRichman> the fixed parameters thing is a bit of a pain
[14:24] <Randomskk> but is probably something we want to keep anyway since ours do tend to be quite fixed
[14:24] <Randomskk> so, both
[14:24] <Randomskk> predictor tab can show the current 120 hour predictions as a grouped list of tracks
[14:25] <Randomskk> can generate a single prediction for one start time
[14:25] <eroomde> I know that. but to be useful somewhere in there there should be a 120hr landing spot feature. if there's the resource overhead to make it more flexible, excellent
[14:25] <Randomskk> or can generate N predictions at M interval (at most x total tracks) as a group of tracks
[14:25] <eroomde> but this would be the bare minimum
[14:25] <DanielRichman> well if it only takes 20 seconds it might be extra effort have it run on cron. As long as we implement some sort of caching then there's not much advantage
[14:25] <Randomskk> having all three isn't hard and means we can very quickly check the 120 for standard ones
[14:26] <Randomskk> potentially authorised users (or just like, two or three of us) allowed to have a scenario that generates the 120hr ones automatically
[14:26] <Randomskk> and let anyone generate fewer predictions at whatever params and intervals they like
[14:26] <DanielRichman> We've got two cores here also; is the predictor parlellised in any way?
[14:26] <DanielRichman> 'cause you could cut that down to 10 seconds by running half of them on the other core
[14:26] <Randomskk> not usefully but the hourly-predictions runs separate predictions in multiple threads
[14:26] <Randomskk> 20 seconds is running on the 8? cored SRCF machine
[14:26] <Randomskk> >.>
[14:26] <DanielRichman> oh.
[14:27] Action: Randomskk hacked the threading in a few months back, it made a massive difference
[14:27] <DanielRichman> heh what did the SRCF think about you hogging every single cycle for 20s?
[14:28] <DanielRichman> btw Randomskk I didn't get around to finishing monit's config; see /etc/monit.conf later on if you have a moment. I think I commented out the mysql block to disable it
[14:28] <Randomskk> they seem happy
[14:28] <Randomskk> well I say that
[14:28] <Randomskk> eroomde: we need to clear some files
[14:28] <Randomskk> we are hitting our quota again and it's already bigger than the usual quota
[14:29] <Randomskk> SRCF said they'll increase it but politly asked for a donation
[14:31] <Randomskk> it's already much bigger as I understand it
[14:31] <Randomskk> 2x or so
[14:31] <Randomskk> mostly our svn is friggin massive and there are a load of gigantic images
[14:33] <DanielRichman> How often are GFS files released?
[14:34] <Randomskk> 4 times a day
[14:34] <Randomskk> datasets 00z, 06z, 12z and 18z and are released about 5 to 6 hours after that time
[14:34] <Randomskk> ish
[14:35] <Randomskk> http://nomads.ncep.noaa.gov:9090/dods/gfs_hd/gfs_hd20100806
[14:35] <Randomskk> make that 4 to 5 hours
[14:35] <Randomskk> so no point running more frequently than that
[14:36] <Randomskk> srcf cronjon runs too often last I checked, jonsowman might have fixed it
[14:36] Action: DanielRichman is unfamiliar with the gfsness
[14:36] <Randomskk> his hourly predictor runs at 0500 and every six hours thereafter iirc
[14:36] <DanielRichman> so how large a timespan does that data cover?
[14:36] <Randomskk> each one of those datasets is what, a week
[14:36] <Randomskk> and they release a new one every six hours
[14:36] <Randomskk> but the python script just downloads the bits required
[14:36] <Randomskk> http://nomads.ncep.noaa.gov:9090/dods/gfs_hd/gfs_hd20100806/gfs_hd_06z.info
[14:36] <Randomskk> that is the 6AM UTC today dataset
[14:37] <Randomskk> covers 0 to 359.5 long, -90 to 90 lat, 1000 to 1 alt, 06Z06AUG2010 to 06Z14AUG2010 time
[14:37] <DanielRichman> OK. How does it download just the bits required; is there a separate file for smaller sections?
[14:37] <Randomskk> no
[14:37] <Randomskk> there's no 'file'
[14:37] <Randomskk> it's accessed via the DAP protocol
[14:37] <Randomskk> python uses the PyDAP library
[14:37] <DanielRichman> Right
[14:37] <Randomskk> look at all that lovely data. surface albedo for any spot on earth
[14:37] <Randomskk> convective precip rate
[14:38] <Randomskk> categorical rain for any part of earth, predicted
[14:38] <Randomskk> kg/m² cloud water
[14:38] <Randomskk> temperatures at various heights above sea and ground
[14:38] <DanielRichman> and how big in GB is data for the whole world?
[14:38] <Randomskk> kg/m² surface water runoff
[14:38] <Randomskk> no idea. shittons.
[14:38] <Randomskk> you don't do that
[14:38] <Randomskk> besides, data what
[14:39] <eroomde> sorry was talking to boss - Randomskk whassat?
[14:39] <Randomskk> this dataset has everything
[14:39] <Randomskk> eroomde: the GFS data set
[14:39] <Randomskk> http://nomads.ncep.noaa.gov:9090/dods/gfs_hd/gfs_hd20100806/gfs_hd_06z.info
[14:39] <eroomde> yes we reach our data limit periodically
[14:39] <eroomde> i deleted a few hundred MB a few months back wehn it stopped working
[14:40] <Randomskk> DanielRichman: we just use hgtprs, ugrdprs, vgrdprs data
[14:40] <DanielRichman> I C
[14:40] <Randomskk> and we just get it for a relatively small time period and the UK lat/lon
[14:40] <DanielRichman> so is there any point downloading hgtprs ugrdprs vgrdprs data for the whole of the UK when its released and holding it on the HDD?
[14:40] <Randomskk> for what time scale?
[14:41] <Randomskk> kind of, yes, in that it is cacheable and makes things much faster
[14:41] <Randomskk> takes otoh hundreds of megs so space isn't an issue, bandwidth might be
[14:41] <Randomskk> dunno if Upu is unmetered or what
[14:41] <Randomskk> also it takes ages, download rate is not all that fast
[14:41] <DanielRichman> ic
[14:41] <Randomskk> hence current method of downloading as needed and caching until there is a newer dataset
[14:42] <Randomskk> right now it could still download less though
[14:42] <DanielRichman> yes, we download the whole square, don't we?
[14:42] <Randomskk> if the predictor could fetch data just as the balloon needed it...
[14:42] <DanielRichman> indeed.
[14:42] <Randomskk> that would require a C++ DAP client or a Python predictor (or a lot more interprocess communciations) though
[14:42] <Randomskk> atm the python also converts the gfs to a csv file for the predictor, predictor doesn't understand gfs
[14:42] <DanielRichman> C++ DAP clients must be out there
[14:42] <Randomskk> no doubt
[14:43] <Randomskk> a C-based pythonic predictor could probably also work, depending on if you did the monte carlo predictions or normal
[14:43] <Randomskk> where normal is faster and easier to write
[14:43] <eroomde> yeah it would make the 120hr prediction about 400 times slower though
[14:44] <Randomskk> well probably not but it would definitely make it slower
[14:44] <Randomskk> if we time the hourly preds right they should download all the data anyone is likely to need
[14:44] <Randomskk> and thus it'l all be cached
[14:44] <Randomskk> for the UK anyway
[14:45] <DanielRichman> if we gave the predictor a DAP client and had it only download the data it needed, we could create a intermediary "DAP-cache" server
[14:45] <DanielRichman> if such a thing exists
[14:45] <DanielRichman> squid for dap
[14:45] <Randomskk> well it's kinda
[14:45] <Randomskk> that's kinda what it does
[14:45] <Randomskk> kinda
[14:45] <Randomskk> pydap has a cache folder
[14:45] <Randomskk> it caches http requests
[14:46] <Randomskk> via httplib or whatever it's called
[14:46] <Randomskk> httplib2 caches
[14:46] <eroomde> httplib2 nowadays
[14:46] <Randomskk> it caches http transactions
[14:46] <eroomde> gotta keep up with the joneses'
[14:46] <Randomskk> pydap uses http
[14:46] <DanielRichman> lovely
[14:46] <Randomskk> however http://github.com/jonsowman/cusf-standalone-predictor/blob/master/predict.py#L22
[14:46] <Randomskk> the data back from the nomads server doesn't include caching information
[14:46] <Randomskk> so libhttp2 doesn't cache it, normally
[14:47] <Randomskk> so predict.py is monkeypatched fantastically to just tell libhttp2 to use the cached data
[14:47] <DanielRichman> :)
[14:47] <Randomskk> unless the server said not to cache it explicitely, which it does in the case of non-existing datasets
[14:47] <Randomskk> so we don't cache the non-existance of a more recent dataset (the script checks for 18z, 12z, 6z, 0z for the current day)
[14:47] <Randomskk> (in that order)
[14:47] <Randomskk> but we do cache all the other data, and force libhttp2 to use it
[14:48] <Randomskk> thus anything predict.py downloads gets cached to the httplib2 cache directory
[14:49] <Randomskk> if hourly-predictor was suitably monkeypatched, or better yet the GFS getting was abstracted out to something else, they'd all use the same cache files where possible
[14:49] <DanielRichman> and that's multiprocess safe?
[14:49] <Randomskk> maybe
[14:50] <Randomskk> and so the hourly predictor run every six hours would fetch all the data it's possible to get anyway (for the UK) and thus all predictions for the next six hours inside the UK would be using cached data and thus very fast
[14:50] <eroomde> DanielRichman: to answer your question about multi-threaded prediction, there is a CUDA one which does 600 predictions in the time it currently takes to do one
[14:50] <DanielRichman> :D
[14:50] <eroomde> but I don't suppose this box has a GPU
[14:50] <Randomskk> how many predictions does it require for a good monte carlo result though?
[14:50] <eroomde> poor wee nessie
[14:50] <Randomskk> 600? :P
[14:50] <eroomde> um, you could fill several PhD theses answering that question
[14:50] <eroomde> but yes, 600 is fine
[14:51] <Randomskk> talking of which, have you made your cambridge:australia decision?
[14:52] <eroomde> no
[14:54] <Randomskk> :o choices choices
[15:01] <DanielRichman> I guess I ought to email the list :)
[15:03] <Randomskk> yea, summarise all that design detail
[15:03] <Randomskk> I want to do something with all this other exciting gfs data
[15:03] <Randomskk> look at it, there's everything there
[15:04] <Randomskk> weasdsfc
[15:04] <Randomskk> ** surface water equivalent of accumulated snow depth [kg/m^2]
[15:04] LazyLeopard (~irc-clien@chocky.demon.co.uk) joined #highaltitude.
[15:04] <Randomskk> how they predict all this I have no idea
[15:09] <SpeedEvil> I think it's done with a massive array of seaweed.
[15:09] jasonb_ (~jasonb@adsl-66-124-73-250.dsl.sntc01.pacbell.net) left irc: Quit: Leaving
[15:16] jasonb (~jasonb@adsl-66-124-73-250.dsl.sntc01.pacbell.net) joined #highaltitude.
[15:29] <DanielRichman> Randomskk; eroomde; is this accurate? http://pastebin.com/zmMtVeGD
[15:30] <eroomde> yes
[15:32] <eroomde> from my perspective anyway
[15:33] <DanielRichman> ok. I'll send it then
[15:34] <DanielRichman> I think it will be ok
[15:36] <DanielRichman> Randomskk: making the predictor request data looks quite easy using a C DAP library
[15:36] <DanielRichman> just need to replace the get_wind(...) function
[15:36] <DanielRichman> I might have a go
[16:13] jasonb (~jasonb@adsl-66-124-73-250.dsl.sntc01.pacbell.net) left irc: Ping timeout: 265 seconds
[16:25] jasonb (~jasonb@m390536d0.tmodns.net) joined #highaltitude.
[16:30] <Randomskk> famous last words, but you are clearly quite capable, go for it :P
[16:56] GW8RAK (~chatzilla@client-86-31-213-22.oxfd.adsl.virginmedia.com) joined #highaltitude.
[17:27] jasonb (~jasonb@m390536d0.tmodns.net) left irc: Ping timeout: 276 seconds
[17:40] jasonb (~jasonb@dsl092-009-225.sfo1.dsl.speakeasy.net) joined #highaltitude.
[17:57] <griffonbot> @nearsys: Just got word, the PCBs are on their way back from China. #ARHAB [http://twitter.com/nearsys/status/20483729711]
[17:59] <DanielRichman> ping Randomskk
[17:59] juxta (juxta@ppp203-122-193-94.static.internode.on.net) joined #highaltitude.
[18:05] jasonb (~jasonb@dsl092-009-225.sfo1.dsl.speakeasy.net) left irc: Ping timeout: 245 seconds
[18:09] jasonb (~jasonb@dsl092-009-225.sfo1.dsl.speakeasy.net) joined #highaltitude.
[18:15] <Randomskk> DanielRichman: hi
[18:15] <Randomskk> sorry was at a free bbq :D
[18:15] <DanielRichman> :O
[18:16] <Randomskk> apparently the technology centre does them once a year
[18:16] <DanielRichman> nah don't worry; I think I've found the answer to my question
[18:16] <Randomskk> so, good timing
[18:16] <Randomskk> want to ask anyway or are you all okay?
[18:16] <DanielRichman> I think I'm all ok
[18:16] <Randomskk> cool
[18:24] DaveyC (~IceChat7@188-221-51-13.zone12.bethere.co.uk) joined #highaltitude.
[18:28] DaveyC (~IceChat7@188-221-51-13.zone12.bethere.co.uk) left irc: Client Quit
[18:29] juxta (juxta@ppp203-122-193-94.static.internode.on.net) left irc: Ping timeout: 260 seconds
[18:36] GW8RAK (~chatzilla@client-86-31-213-22.oxfd.adsl.virginmedia.com) left irc: Remote host closed the connection
[18:50] Hiena (~Hiena@ joined #highaltitude.
[18:50] <fsphil> I'm never helping someone put down concrete again .. that's hard work!
[18:55] <DanielRichman> Randomskk: I'm not 100% sure that it would be worth it; the penalty from having it make multiple requests to the DAP server must be quite high
[18:59] <Randomskk> back
[18:59] <Randomskk> DanielRichman: hard to say. it's a balacing act of how many requests vs how much data each
[18:59] <Randomskk> I'm not convinced what we have now is optimum, but it is easiest, especially if we just cache everything
[18:59] <Randomskk> yay, back home and it's the weekend
[18:59] <Randomskk> thank god for that
[19:01] <DanielRichman> oh yeah; it is the weekend >.>
[19:01] <Randomskk> you bastards and your "school holidays"
[19:01] <Randomskk> remind me what A levels you're doing?
[19:01] <Randomskk> doing/will do
[19:02] <DanielRichman> maths furthermaths chemistry physics german
[19:02] <DanielRichman> Randomskk: I think I'll pass up trying to do some DAP in C just for now. Maybe another time
[19:02] <Randomskk> fair enough
[19:02] <Randomskk> no DT/electronics?
[19:03] <DanielRichman> Did elec for gcse. Not sure whether I should do it for A level
[19:03] <Randomskk> planning on doing chemistry for A level, or just AS?
[19:03] <DanielRichman> Chemistry the whole way
[19:03] <Randomskk> crazy. you'll regret that
[19:03] <DanielRichman> Chemistry department at our school is very good
[19:03] <Randomskk> DT: systems for A level was so much fun for me, my teacher just gave me a silly budget and I made a robot, it was easily the most relaxing and fun lesson of the week
[19:03] <Randomskk> (mainly because it wasn't a lesson, just a case of playing with electronics)
[19:03] <DanielRichman> besides; I can always bail out at y12
[19:04] <Randomskk> fair enough
[19:04] <DanielRichman> Randomskk: Elec gcse class size was three
[19:04] <DanielRichman> there were a couple of occasions where I'd get a ""lesson"" on my own
[19:04] <DanielRichman> etc.
[19:04] <Randomskk> do I know three of them? >_>
[19:04] <Randomskk> my DT A level was 5, but eh
[19:04] <DanielRichman> mmm dt a level does sound like fun. I guess I just decided to do Gm instead. What else did you do?
[19:06] <Randomskk> maths, further maths, physics, critical thinking for AS
[19:06] <Randomskk> we had to do crit. thinking of philosophy for AS
[19:06] <Randomskk> I really enjoyed critical thinking, I think I was the only person in my year to do so, everyone else thought it was shit
[19:06] <Randomskk> everyone else sucks
[19:06] <DanielRichman> :)
[19:07] <DanielRichman> our school offers ... I forget what it's called... "theory of knowledge" or smth. but AlexBreton & ssb & a few others have been lobbying them to start teaching phiolosophy
[19:07] <DanielRichman> neither are compulsory though
[19:08] <Randomskk> fair enough
[19:08] <Randomskk> the thing about my DT was I could do anything I wanted as my project, literally
[19:08] shauno (soneil@pdpc/supporter/professional/shauno) left #highaltitude.
[19:08] <Randomskk> so long as I could write about it and it worked, I was free to pick anything
[19:09] <Randomskk> if I'd have had to do something boring it'd have been much worse
[19:09] <Randomskk> or I'd have finished it in one term and had two terms of frees
[19:09] <DanielRichman> :P
[19:09] <Randomskk> the robot took way more time and effort
[19:09] <DanielRichman> I saw the pics/docs of that, it looked pretty cool :P
[19:10] <Randomskk> it worked, which was a pleasant surprise
[19:10] <DanielRichman> I've seen Air Hockey tables go past in our tech department, from the a level students
[19:10] <DanielRichman> :)
[19:10] <Randomskk> nice
[19:10] <DanielRichman> <-- bbl; eating
[19:10] <Randomskk> that would be a lot of fun
[19:10] GW8RAK (~chatzilla@ joined #highaltitude.
[19:21] LazyLeopard (~irc-clien@chocky.demon.co.uk) left irc: Quit: Bye
[19:30] SiOB (~SiOB@cpc7-stkp7-2-0-cust544.know.cable.virginmedia.com) joined #highaltitude.
[19:42] <DanielRichman> back.
[19:42] <Randomskk> yo
[19:44] <DanielRichman> monit keeps on complaining about exim
[19:44] <Randomskk> hmm
[19:44] <Randomskk> not sure what it counts as a failure
[19:48] <DanielRichman> Aug 6 18:47:53 nessie monit[13192]: SMTP: error receiving data -- Resource temporarily unavailable
[19:48] <DanielRichman> Aug 6 18:47:53 nessie monit[13192]: 'exim' failed protocol test [SMTP] at INET[localhost:25] via TCP
[19:51] <Randomskk> curious
[19:51] <Randomskk> oh my, `tree`
[19:52] <DanielRichman> hmm?
[19:52] <Randomskk> I hadn't come across tree before
[19:53] <DanielRichman> it's one of the first things I install
[19:53] <DanielRichman> along with pv
[19:56] Daviey (~Daviey@ubuntu/member/daviey) left irc: Excess Flood
[19:56] <Randomskk> is there anything to do on the server atm?
[19:57] <DanielRichman> unless you fancy getting monit to <3 mysql
[19:57] <Randomskk> monit doesn't <3 mysql?
[19:57] <DanielRichman> :'( noo
[19:57] <DanielRichman> # TODO: Fix conf.d/mysql . Perhaps enable monitoring for the php fcgi processes?
[19:57] <DanielRichman> (-- monitrc:2)
[19:58] <DanielRichman> although you may want to change the email monit spams before doing so :P
[19:58] <DanielRichman> I don't mind, but some people might
[20:00] Daviey (~Daviey@ubuntu/member/daviey) joined #highaltitude.
[20:04] <Randomskk> unless you did, I didn't do anything there
[20:04] <Randomskk> monit should be up and working with mysql though
[20:05] <DanielRichman> ok
[20:05] <Randomskk> I just killed mysql
[20:05] <Randomskk> but my last monit email just says monit stopped, so not sure if it was actually monitoring or not
[20:06] <Randomskk> basically mysql was not configured to make a pidfile
[20:06] <Randomskk> now it is
[20:06] <Randomskk> okay, monit got mysql stopping
[20:06] <Randomskk> and mysql is running again
[20:07] <DanielRichman> cool. So did you make any changes?
[20:07] <Randomskk> yea
[20:07] <Randomskk> they are committed
[20:08] <DanielRichman> oh I see
[20:10] <Randomskk> http://captchart.posterous.com/captchart haha awesome
[20:12] <Randomskk> context: 4chan apparently now has a captcha, I'm told
[20:14] <DanielRichman> sure you are
[20:14] Action: Randomskk hasn't been on 4chan in coming on a year
[20:14] <DanielRichman> anyway rjh just called; he was wondering if we should install the current predictor & tracker on habhub.org so that we can start directing people there, or if we should save it for the "big release"
[20:15] <Randomskk> eh
[20:15] <Randomskk> we could do
[20:15] <Randomskk> I think hacking on it first is a good idea
[20:15] <Randomskk> depends how much work there is to move and set everything up
[20:15] <Randomskk> given as a fair bit is probably going to be changed?
[20:16] <DanielRichman> well setting up predictor v2 is very easy but the tracker might be harder to shift
[20:16] <Randomskk> yea exactly
[20:16] <DanielRichman> and the predictor v2 is already in 2 places
[20:16] <Randomskk> also I want python stuff Done Right w.r.t. packages, virtualenv, etc
[20:16] <DanielRichman> hrrm I dunno how much python we're going to end up with at this rate
[20:17] <Randomskk> in other words avoiding installing everything systemwide for everything half from apt and half from easy_install and half from that other pythonic one
[20:17] <DanielRichman> btw; is there an easy_uninstall?
[20:17] <Randomskk> who knows
[20:17] <Randomskk> easy_install is now 'not cool'
[20:17] <Randomskk> the cool kids are using, uh
[20:17] <Randomskk> what is it
[20:17] <Randomskk> pip?
[20:18] <DanielRichman> wat.
[20:18] <Randomskk> distutils2?
[20:18] <DanielRichman> get out.
[20:18] <Randomskk> I think it's pip
[20:18] <DanielRichman> okey.
[20:18] <Randomskk> for instance, http://clemesha.org/blog/2009/jul/05/modern-python-hacker-tools-virtualenv-fabric-pip/
[20:19] <DanielRichman> riight.
[20:19] <DanielRichman> I want an easy_install that converts the silly python packages to debs
[20:19] <DanielRichman> then installs using dpkg. I trust dpkg
[20:25] <Randomskk> yea
[20:25] <Randomskk> but again that leads to system wide
[20:25] <Randomskk> I'm becoming more and more torn about that
[20:25] <Randomskk> http://thewikigame.com/ is quite amusing
[20:26] <Randomskk> on the one hand, no duplication
[20:26] <Randomskk> on the other hand, python and ruby etc libraries are so frequently updated...
[20:28] <DanielRichman> haha I didn't realise there was a dedicated site for the wiki game
[20:28] <DanielRichman> 'spose it wasn't going to be long before there was one, was it
[20:28] <Randomskk> nor did I until just now
[20:31] <DanielRichman> Randomskk: are you playing?
[20:31] <Randomskk> I was
[20:31] <Randomskk> I am not at the moment
[20:31] <DanielRichman> hehe
[20:33] <DanielRichman> hmm. OpenNDAP sees that you have a C applicatino you want to use with DAP and gives you two options (how kind): 1) have this lovely documented c++ library and enjoy linker/compiler hell, or 2) have this totally documentationless C library
[20:36] <Randomskk> hey so
[20:36] <Randomskk> open street map
[20:36] <Randomskk> you can download the entire map
[20:36] <Randomskk> it'l fit on a DVD
[20:37] <Randomskk> I guess everyone knew this and this was the whole point
[20:37] <Randomskk> but anyway for a hab live cd or such it would make maps work quite well without any google hackery
[20:37] <DanielRichman> ed was talking about Ordanance Survey maps
[20:37] <Randomskk> so he was
[20:37] <DanielRichman> 'parrently free for noncommercial
[20:37] <Randomskk> though the osm ones seem to have footpaths too
[20:37] <Randomskk> however yes, OS would be Even Better
[20:38] <Randomskk> the entire worldwide osm data is Quite Big
[20:38] <DanielRichman> I didn't realise the whole of OSM would fit onto a dvd
[20:38] <DanielRichman> is that in png format or their funky XML format?
[20:38] <DanielRichman> ie rendered or not
[20:39] <Randomskk> well
[20:39] <SpeedEvil> png it's around a terabyte
[20:39] <SpeedEvil> or several
[20:39] <Randomskk> fcvo fit
[20:39] <SpeedEvil> I forget the exact numbers.
[20:39] <Randomskk> they say about 7GB with bzip2
[20:39] <Randomskk> it extracts to Quite Big
[20:39] <SpeedEvil> that's the xml
[20:39] <SpeedEvil> 160G IIRc
[20:39] <Randomskk> yes
[20:39] <Randomskk> so perhaps worldwide would be silly
[20:40] <SpeedEvil> though you can I think sanely fit it in a compact binary representation in RAM in ~12G
[20:40] SiOB (~SiOB@cpc7-stkp7-2-0-cust544.know.cable.virginmedia.com) left irc:
[20:41] <DanielRichman> Randomskk: you removed `finger` but you forgot about `pinky`
[20:41] <Randomskk> haha
[20:42] <DanielRichman> daniel@nessie:~$ dpkg -S `which pinky`
[20:42] <DanielRichman> coreutils: /usr/bin/pinky
[20:42] <DanielRichman> ^^ you're not removing that >.>
[20:42] <Randomskk> :P
[20:43] <DanielRichman> I wonder what happens if you put terminal escape codes in .plan ....
[20:43] <Randomskk> bad things used to happen
[20:44] <DanielRichman> lul `pinky -l daniel` enjoy
[20:44] <Randomskk> excellent
[20:44] <DanielRichman> 6.4MB of /dev/urandom faw juu
[20:52] LA3QMA (~kgb@ joined #highaltitude.
[21:01] tydeas (~Energy@ joined #highaltitude.
[21:09] <DanielRichman> my bad; the C only library *was* documented, but the documentation could only be found in SVN, not their source tarball
[21:10] rjharrison (~rharrison@ joined #highaltitude.
[21:10] <rjharrison> evening all
[21:10] <Randomskk> hi
[21:10] <rjharrison> I have a five minute window to come and play for a bit
[21:11] <rjharrison> I seem to have a new spammer in my inbox called monit
[21:11] <Randomskk> hehe
[21:12] <rjharrison> I guess she'll calm down once things settle on the server
[21:12] <Randomskk> yup, should be pretty quiet from now on unless anything actually goes wrong
[21:12] <DanielRichman> Yeah sorry I didn't ask totally before putting everyone down to receive root's mail. If you don't want that, give a shout
[21:13] <rjharrison> I was having a chat with DanielRichman earlier on the phone. I wondered if we wanted to integrate the exising tracker and listener and the onto the server
[21:13] <DanielRichman> rjharrison: Randomskk isn't sure about immediatly putting the current software onto habhub.org
[21:14] <rjharrison> Sure no problem. Is the plan to build the system from scratch and repeat what we have done with a different backend?
[21:14] <Randomskk> I think as much reuse as practical
[21:14] <Randomskk> we need to set out some goals and user stories and specs and things
[21:14] <Randomskk> nothing too complicated obviously, no point wasting time on it
[21:15] <Randomskk> but enought to know what we want to do
[21:16] <rjharrison> Sure but there may be some sense in getting the exisitng system working under one roof and setting up the clients to send the data to habhub and the exisiting system stays in place untill the new is ready.
[21:16] <rjharrison> I'm not that fussed given the system sort of works now.
[21:18] <Upu> evening
[21:18] <rjharrison> But it might be worth capitalising on the momentum for centralisation
[21:18] <Randomskk> hi
[21:18] <rjharrison> hi Upu
[21:19] <Upu> yeah monit is definitely active :)
[21:19] <Randomskk> should calm down now hopefully
[21:19] <Upu> I know there is an issue as my iphone pings and the dog barks
[21:19] <rjharrison> lol
[21:24] <Upu> I hate ISP some times :
[21:24] <Upu> We are aware of a small number of 20CN and 21CN DSL customers who are experiencing sporadic performance and connectivity related problems with both VoIP and IPSEC traffic.
[21:24] <Upu> by small number they mean all
[21:24] <Upu> by sporadic they mean all the time
[21:24] <Upu> and by performance they mean doesn't work at all
[21:26] <Randomskk> :D
[21:27] <Randomskk> Upu: by the way, is the server metered? as in, does it make a difference to you how much bandwidth it uses?
[21:29] <Randomskk> brb dinner
[21:32] <Upu> don't worry about bandwidth
[21:32] tydeas (~Energy@ left irc:
[21:34] <Randomskk> cool, okay
[21:34] <Randomskk> waitrose chicken vindaloo is okay but not very spicy, but with the addition of nandos extra extra hot sauce it is excellent
[21:36] <DanielRichman> :D nando's sauce
[21:37] <fsphil> hot is good
[21:37] <Randomskk> :D yup
[21:37] <Randomskk> I think nandos spice is my favourite kind of spice
[21:37] <Randomskk> followed by indian?
[21:38] <Randomskk> not as much a fan of ginger or wasabi
[21:38] <Randomskk> though wasabi is excellent in sushi, of course
[21:38] <fsphil> I always suffer after eating hot indian food, but I keep going back.
[21:38] <sbasuita> the extra extra hot nandos sauce isn't hot enough...
[21:38] <Randomskk> sbasuita: I know right
[21:39] <Randomskk> but it's the best I can find in waitrose
[21:39] <Randomskk> and it's okay for everyday stuff
[21:39] <Randomskk> also it is tasty
[21:39] <sbasuita> Randomskk: yeah i'm not really a fan of ridiculous spice anyway
[21:39] <sbasuita> its just they put nuclear warnings and shit all over it for some reason...
[21:40] <sbasuita> (at least in the restaurants)
[21:40] <Randomskk> in the restaurants it's a bit silly
[21:40] <Randomskk> though tbh I find their very hot is as hot or hotter than the very very or whatever it's called
[21:41] <fsphil> meh, as long as they don't serve it in a hazmat suit
[21:41] <sbasuita> haven't been to nandos in ages :(
[21:41] <Randomskk> :(
[21:41] <sbasuita> remember some good times tehre
[21:41] <Randomskk> nor have I actually
[21:41] <Randomskk> there is one in cambridge but I have only been there once
[21:43] <DanielRichman> sometimes I put that stuff in sandwiches.
[21:43] <Randomskk> it is far and expensive, the curry house is down the road and cheaper and does thaals
[21:43] <Randomskk> nomnomnom
[21:44] <DanielRichman> hmm I'm getting the feeling that it's easier to write a dap library in an interpreted language
[21:45] <Randomskk> so: make a pythonic predictor :P
[21:45] <DanielRichman> that'd be super slow
[21:45] <Randomskk> with C extensions
[21:45] Hiena (~Hiena@ left irc: Quit: -=Got bored from the net. Gone blowing up things.=-
[21:45] <DanielRichman> You could launch and recover the balloon faster than run a prediction
[21:46] <Randomskk> bull
[21:46] <Randomskk> python's not that much slower
[21:46] <DanielRichman> python with C extensions? hmm. Depends on how many times the DAP download needs to be hit
[21:46] <DanielRichman> I think I'd go for C with python extensions >.>
[21:46] <DanielRichman> I suppose the python libs are better documented but that's yet *another* set of docs to wade through. OpeNDAP's documentation is unimpressive
[21:47] <DanielRichman> all I wanted was an example
[21:48] <Randomskk> and now a white choc magnum to cool off
[21:48] <Randomskk> <3
[21:48] <Randomskk> my parents are out so I have to fend for myself
[21:48] <Randomskk> god I'd die if I ever had to cook for myself on a regular basis
[21:48] <Upu> So waitrose curry and a choc magnum
[21:48] <DanielRichman> I haven't had an ice cream in like... ages... at least a week
[21:49] <Randomskk> Upu: and some rice with the curry!
[21:49] <Upu> lol
[21:49] <Randomskk> and some fruit beforehand!
[21:49] <Upu> oh that makes it all ok then :)
[21:49] <Randomskk> vegetables are way more effort than they're worth though
[21:51] <Randomskk> right well I was going to document what all the PHP in the predictor does with a view of pythonising it, but now I'm tempted to make a proof of concept python predictor
[21:54] <DanielRichman> Randomskk: well I've only just about got my head around the C predictor
[21:54] <DanielRichman> to teh extent that I'm about to draw a picture
[21:54] <Randomskk> I haven't even started to look at it
[21:55] <DanielRichman> so if I finish the picture I'll upload it; might be useful
[21:55] <Randomskk> I understand it's relatively complicated
[21:55] <Randomskk> that'd be good
[21:55] <DanielRichman> however I am going to try and do this DAP-C thing
[21:55] <DanielRichman> I think
[21:56] <Randomskk> okay
[21:56] <DanielRichman> however Randomskk I'm ignoring absolutely everything apart from the get_wind function
[21:56] GW8RAK (~chatzilla@ left irc: Remote host closed the connection
[21:57] <DanielRichman> it looks like it was started with the best intentions but a few features added afterwards; you know how it is
[21:58] <DanielRichman> tonnes of stuff is passed between functions as you'd expect; but `FILE *kml_file` is a global
[21:58] <Randomskk> lovely
[21:58] <Randomskk> I think its KML outputting is broken
[21:58] <Randomskk> at least, it's always segfaulted every time I tried
[21:58] <Randomskk> the predictor has a long and dark history
[21:58] <Randomskk> there's original CPP for it around somewhere
[22:01] <DanielRichman> what.
[22:01] <DanielRichman> why
[22:01] <Randomskk> see: long, dark
[22:01] <DanielRichman> I won't go there
[22:01] <Randomskk> this one dates to 07/08, hmm
[22:02] Action: DanielRichman runs git log
[22:02] <DanielRichman> hmm stonewall from import from svn
[22:02] <Randomskk> there was an svn import?
[22:02] <Randomskk> at any rate good, the svn is full of shit
[22:03] <Randomskk> it's hundreds of megs of random crap
[22:03] <Randomskk> http://www.srcf.ucam.org/~cuspaceflight/websvn/filedetails.php?repname=CUSF&path=/landing_prediction/pre_flight_prediction/landing_prediction.cpp
[22:03] <Randomskk> that's one of the originals
[22:04] <Randomskk> data = new float****[use_Nj+1];
[22:04] <Randomskk> sexy
[22:04] <Randomskk> it worked, though
[22:06] <DanielRichman> oh lovely.
[22:07] <DanielRichman> altitude_model_get_altitude(self, time, *alt)
[22:07] DaveyC (~IceChat7@188-221-51-13.zone12.bethere.co.uk) joined #highaltitude.
[22:07] <DanielRichman> *alt is an output in all cases except for the first time you call the function, in which case it's an input for the initial altitude (despite ther e being a perfectly good altitude_model_new)
[22:07] <DanielRichman> not that I want to disrespect the work that's gone into this
[22:10] <Randomskk> and it does work
[22:10] <Randomskk> as with any code, it just looks horrific when you see all the hacks in the light of day
[22:11] <DanielRichman> :)
[22:11] <DanielRichman> and it is fairly well documented
[22:11] <DanielRichman> there's comments and the variables/functions have good enough names that I can sort of work out whats going on
[22:12] <DanielRichman> PS altitude module get altitude; time argument is totally ignored on descent; it assumes a hardcoded TIMESTEP
[22:20] <Randomskk> excellent
[22:21] <DanielRichman> I guess the question is, if it is going to be rewritten, should it be in C or Python? (actually I reckon it only needs a bit of tidying in C)
[22:21] <DanielRichman> perhaps a "predictor_job" struct rather than passing all these stack variables to each function recursivly
[22:22] <Randomskk> or even C++
[22:22] <DanielRichman> mmm. I'm not sure I really lice C++
[22:23] <DanielRichman> *like
[22:23] <DanielRichman> Also there's vague support in here for running multiple predictions at once
[22:23] <Randomskk> it's not that bad really. it has a lot of relatively neat things
[22:23] <Randomskk> I think there is code in existance that does the massively parallel prediction for monte carlo
[22:23] <Randomskk> both on CPU and GPU
[22:23] <Randomskk> but, where?
[22:24] <Randomskk> man, oscopes. 2 channel, 25mhz: £270 40mhz: £305 60mhz: £350 100mhz: £500 150mhz: £660 200mhz: £800
[22:24] <Randomskk> I'm torn between the 40 and 60 I think
[22:25] <Randomskk> 60 is like £45 more for like 20mhz more
[22:25] <Randomskk> but potentially quite useful 20mhz
[22:29] <DanielRichman> Hmm perhaps you could speed up the monte carlo by not using qsort to pick the "most likely state"
[22:29] <Randomskk> nessie has an ATI Rage XL
[22:30] <Randomskk> probably not great for GPU maths :P
[22:31] <DanielRichman> And I doubt the chase car will be ready to pull of large CUDA jobs either
[22:31] <Randomskk> yea
[22:32] <DanielRichman> 'specially on an ATI card
[22:34] SpeedEvil (1000@tor/regular/SpeedEvil) left irc: Ping timeout: 276 seconds
[22:45] G8DSU (~chatzilla@cpc3-mort4-0-0-cust90.croy.cable.virginmedia.com) joined #highaltitude.
[22:46] LazyLeopard (~irc-clien@chocky.demon.co.uk) joined #highaltitude.
[22:49] SpeedEvil (~user@tor/regular/SpeedEvil) joined #highaltitude.
[22:59] DaveyC (~IceChat7@188-221-51-13.zone12.bethere.co.uk) left irc: Ping timeout: 252 seconds
[23:04] DaveyC (~IceChat7@188-221-51-13.zone12.bethere.co.uk) joined #highaltitude.
[23:05] DaveyC (~IceChat7@188-221-51-13.zone12.bethere.co.uk) left irc: Client Quit
[23:13] <Randomskk> yay
[23:13] <Randomskk> oscilloscope on its way
[23:14] <Upu> I think JackNorris is getting one too this week ?
[23:14] <Upu> I'm going to order a radio on Monday
[23:14] <Randomskk> :o
[23:14] <Randomskk> I have a lovely radio but it desperately needs a proper power supply
[23:14] <Randomskk> got a tenma 60mhz 2ch scope
[23:14] <Randomskk> it has a colour screen and usb and maths and shit :o
[23:14] <Upu> sounds better than the last one I used
[23:15] DaveyC (~IceChat7@188-221-51-13.zone12.bethere.co.uk) joined #highaltitude.
[23:20] <DanielRichman> Randomskk: http://pastebin.com/fW9xLHmr as I see it, is how it works
[23:21] <DanielRichman> tomorrow I might have a play around with the C, bit of refactoring maybe
[23:21] <DanielRichman> then I'll see how easy it'd be to add libdap in
[23:23] <DanielRichman> but now I'm going to sleep; night
[23:24] <Randomskk> night
[23:26] <JackNorris> yeah sometime next week
[23:27] <Upu> hows the track-otron coming on ?
[23:28] <JackNorris> very well actually
[23:28] <JackNorris> the whole motion thing is solved now ;)
[23:28] <Upu> maths or application of many electronic sensors ?
[23:28] <JackNorris> both :S
[23:29] <JackNorris> Using a gyro to compensate for the acceleration now
[23:29] <Upu> be interested to see if it works on a car roff
[23:29] <Upu> roof
[23:30] SpeedEvil (~user@tor/regular/SpeedEvil) left irc: Ping timeout: 240 seconds
[23:30] <Upu> I'll have to work out how to put something on those rails on my roof
[23:31] <Upu> http://cgi.ebay.co.uk/AUDI-A4-AVANT-ESTATE-01-07-LOCKABLE-ROOF-BARS_W0QQitemZ280366806700QQcmdZViewItem?rvr_id=121073334199&rvr_id=121073334199&cguid=899eb2ae1250a0e205230ce3fc289a9d#ht_1667wt_939
[23:31] <Upu> that was easy :)
[23:32] <JackNorris> cheaper than i thought
[23:32] <Upu> me too
[23:32] <Upu> be interesting to see how much acceleration you can compensate for
[23:39] <JackNorris> yeah
[23:40] rjharrison (~rharrison@ left irc:
[23:45] <Upu> right I'm off - night
[23:45] Upu (~Upu@ubn.upuaut.net) left irc:
[23:47] SpeedEvil (~user@tor/regular/SpeedEvil) joined #highaltitude.
[00:00] --- Sat Aug 7 2010