--- Day changed Tue Aug 06 2013
+15:12 < Callum__> Is the only image processing that needs doing detecting the edge of the can? (and doing measuring how it changes with time etc)
+15:39 < Callum__> hey
+15:39 < james__> what up
+15:39 < sam_moore> Hi
+15:39 < Callum__> Is the only image processing that needs doing detecting the edge of the can? (and doing measuring how it changes with time etc)
+15:40 < sam_moore> Callum__: I think so
+15:40 < james__> As far as i understand it yes
+15:40 < Callum__> plus just booted up ubuntu on laptop. havent touched it since i did C programming a year ago xD
+15:40 < sam_moore> It would also be nice to just have unprocessed images streamed to the GUI
+15:40 < Callum__> yea what i was thinking
+15:41 < Callum__> firstly, i can look into only processing a portion of the image, so we could get away with a larger resolution
+15:41 < sam_moore> With the server side interface, I'm starting to think I'll ditch the apache2 webserver + CGI and integrate a minimal webserver + the lower level stuff into the same program
+15:41 < Callum__> and openCV has a couple of algorithms to detect edges
+15:41 < Callum__> so will likely look into one or 2 of those
+15:41 < sam_moore> Just thinking about all the layers that are involved with the CGI approach... it probably won't turn out easier to code, and it will definitely be slower
+15:42 < sam_moore> Plus, if we really care about performance, the apache2 webserver has a lot of features we don't care about
+15:42 < james__> As long as i can access the raw files etc i should be able to display it via a gui
+15:42 < sam_moore> The main one being it's designed to cope with multiple users well, and we want to configure it to only allow ONE user
+15:43 < sam_moore> Yes, that sounds good
+15:44 < Callum__> responding to who? me or james?
+15:44 < sam_moore> Um... both now
+15:44 < Callum__> haah
+15:45 < sam_moore> Anyway, I've realised that our software will place an upper limit on the quality of the sensors
+15:45 < james__> true
+15:45 < sam_moore> Since we weren't given a requirement on quality, we should aim to get it as high as feasible
+15:45 < sam_moore> But sensors group needs to have some idea of the upper limit that our software will place on quality
+15:45 < Callum__> yea.
+15:46 < james__> Start simple and ramp it up until the hardware breaks?
+15:46 < sam_moore> No point streaming 1000 images a second if the JavaScript GUI can't send requests that fast.
+15:46 < sam_moore> Yep.
+15:46 < sam_moore> I know you're supposed to do task break downs and all that time management stuff, but the fact is we need to experiment with actual code to work out the best implementation
+15:47 < james__> yeah. break it down into code and test code?
+15:48 < sam_moore> Maybe?
+15:49 < sam_moore> Well, we've got a high level idea of the software
+15:49 < sam_moore> The first tasks should be "decide on the best software implementation of each level"
+15:49 < sam_moore> To do that we need to test some minimalistic implementations
+15:50 < Callum__> yea
+15:50 < sam_moore> So, to set a goal for week 2, can it be "Decide on software implementation at each point in the block diagram"
+15:51 < sam_moore> And "Determine upper limit on data transfer rate through software systems"
+15:51 < sam_moore> The second one might take a bit longer.
+15:52 < Callum__> githubs down for maintenance. hahah
+15:52 < sam_moore> That happened like 2 seconds after I submitted a comment on the CGI stuff
+15:53 < sam_moore> For a while it was showing me a Unicorn (?)
+15:53 < Callum__> yea i got that before
+15:53 < Callum__> when trying to access on my laptop
+15:53 < sam_moore> Well... we have the git.ucc server
+15:54 < sam_moore> But I was starting to like the Issues and Comments and all those actual features
+15:54 < Callum__> hmm. doubt it will be down for long. will have to see
+16:10 < Callum__> back up
+18:35 < Rowan> test test
+18:35 < Rowan> ??
+18:43 < Callum__> Hey
+18:56 -github:#mctxuwa_softdev- [MCTX3420] none pushed 2 new commits to master: http://git.io/ZDMvew
+18:56 -github:#mctxuwa_softdev- MCTX3420/master 64b0cd9 Sam Moore: Investigate software for interfacing with hardware via HTTP requests...
+18:56 -github:#mctxuwa_softdev- MCTX3420/master 5afaa23 Sam Moore: Merge branch 'master' of github:szmoore/MCTX3420
+18:57 -github:#mctxuwa_softdev- [MCTX3420] none pushed 1 new commit to master: http://git.io/1nMEMA
+18:57 -github:#mctxuwa_softdev- MCTX3420/master 9579cea Sam Moore: Makefile for webserver test...
+--- Day changed Wed Aug 07 2013
+00:24 -github:#mctxuwa_softdev- [MCTX3420] none pushed 1 new commit to master: http://git.io/5DEPzw
+00:24 -github:#mctxuwa_softdev- MCTX3420/master fe2fc11 Sam Moore: More work on webserver test...
+00:52 -github:#mctxuwa_softdev- [MCTX3420] none pushed 1 new commit to master: http://git.io/r6aFJw
+00:52 -github:#mctxuwa_softdev- MCTX3420/master e615433 Sam Moore: Add necessary HTTP response headers...
+01:00 -github:#mctxuwa_softdev- [MCTX3420] none pushed 1 new commit to master: http://git.io/eZPWcQ
+01:00 -github:#mctxuwa_softdev- MCTX3420/master b101617 Sam Moore: Automatic commit of irc logs
+10:07 -!- Irssi: #mctxuwa_softdev: Total of 1 nicks [0 ops, 0 halfops, 0 voices, 1 normal]
+20:43 -!- Irssi: #mctxuwa_softdev: Total of 2 nicks [0 ops, 0 halfops, 0 voices, 2 normal]
+20:45 -github:#mctxuwa_softdev- [MCTX3420] none pushed 3 new commits to master: http://git.io/Dhb-Fg
+20:45 -github:#mctxuwa_softdev- MCTX3420/master fe2fc11 Sam Moore: More work on webserver test...
+20:45 -github:#mctxuwa_softdev- MCTX3420/master e615433 Sam Moore: Add necessary HTTP response headers...
+20:45 -github:#mctxuwa_softdev- MCTX3420/master b101617 Sam Moore: Automatic commit of irc logs
+20:45 -github:#mctxuwa_softdev- [MCTX3420] none pushed 3 new commits to master: http://git.io/Dhb-Fg
+20:45 -github:#mctxuwa_softdev- MCTX3420/master fe2fc11 Sam Moore: More work on webserver test...
+20:45 -github:#mctxuwa_softdev- MCTX3420/master e615433 Sam Moore: Add necessary HTTP response headers...
+20:45 -github:#mctxuwa_softdev- MCTX3420/master b101617 Sam Moore: Automatic commit of irc logs
+20:46 < sam_moore> Whoops, I was trying to make it less spammy and ended up making it spam us
+20:53 < Callum__> haha nice work
+20:58 < sam_moore> I have a minimal web server in C done
+20:58 < sam_moore> It was surprisingly easy actually
+21:05 -github:#mctxuwa_softdev- [MCTX3420] none pushed 2 new commits to master: http://git.io/Tui0FA
+21:05 -github:#mctxuwa_softdev- MCTX3420/master c1321a7 Sam Moore: Test webserver with minimalist JavaScript...
+21:05 -github:#mctxuwa_softdev- MCTX3420/master 1b2939d Sam Moore: Merge branch 'master' of github:szmoore/MCTX3420...
+21:07 < sam_moore> Hooray, more spam
+21:07 < sam_moore> Maybe I should just turn that off.
+21:35 < Callum__> Maybe, its kinda goods to know when people commit things though, restricted to one line would be good
+--- Day changed Thu Aug 08 2013
+08:59 -!- Callum_ [
[email protected]] has quit ["AndroIRC - Android IRC Client ( http://www.androirc.com )"]
+--- Day changed Sun Aug 11 2013
+14:09 < Callum__> actually so much work to do already.
+14:47 < sam_moore> Yep
+15:15 < Callum__> well iv spent some time thinking over this, finally get around to writing it down. so iv probably spent close to 4 hours or so thinking about it (while doing other things) but really only spent an hour and a half according to diary
+15:25 < sam_moore> We can always flesh out people's reports with graphs and stuff
+15:27 < Callum__> well we doing 1 per person or group?
+15:29 < sam_moore> 1 per person is the safest way to go I think
+15:29 < sam_moore> We can always summarise them
+15:30 < Callum__> well, tbh i think we should do 1 per person and summarise, but not necessarily go for a full page. thing is not sure how they expect us to do 1 page for 5 people AND be detailed
+15:30 < sam_moore> Yep
+15:31 < Callum__> so they might end up making it 1 per person anyway.
+15:31 < Callum__> not very well thought out either way
+15:31 < sam_moore> Let's just submit a multiple page report, with a bit from each person and a summary. If some people have less than a page that's fine.
+15:32 < Callum__> Yea, agreed thats the best way to go about it. Unless they complain about it being too long
+15:32 < sam_moore> That's why we include a summary :P
+15:33 < sam_moore> Although, it might be difficult to make the summary actually summarise things, since everyone's already using pretty concise summaries
+15:33 < sam_moore> Oh well, writing more stuff is safer to start with.
+15:34 < Callum__> hmm. its kind of a pain though, write it all down in your book...oh look now i need to rewrite most of it for a report..
+15:34 < sam_moore> Ah crap, I forgot about the book
+15:34 < Callum__> haha
+15:34 < sam_moore> Yeah... write it all down in the git commit messages, write it all down in the report, write it all down in the diary...
+15:35 < sam_moore> I suppose in the "Real World" you have to do stuff like this
+15:35 < Callum__> yea true
+18:16 < Callum__> Not sure when il get round to writing what i'v done but il do it tonight some time. We finalising it tomorrow and printing it off then?
+18:27 < sam_moore> Yes, I'd like to have it done before the meeting, or at least do it first thing in the meeting
+18:27 < sam_moore> Then there are some things I need to ask everyone about to progress with my part of the code
+18:27 < sam_moore> Since I'm sort of doing the bit in the middle that gets all the other bits to talk to each other
+18:28 < sam_moore> I think it would be good to keep all the raspberry pi code in a single process, running in seperate threads
+18:29 < sam_moore> But to do that we'd need everyone to use the same language
+18:30 < sam_moore> Also we need to start thinking about the structure of the system in a bit more detail
+18:31 < sam_moore> Like: We can't just transfer data straight to a GUI, because the user will want to use the GUI to start the experiment, then go away and come back later to view results (and possibly save them on their own machine)
+18:32 < sam_moore> Also if we do transfer everything straight to the GUI we'd be limited by the speed of the GUI, and JavaScript requests are slow
+18:34 < Callum__> yea
+18:34 < sam_moore> Pretty easy to just dump stuff to a data file, but we probably also want the ability to do it in realtime
+18:35 < sam_moore> Anyway, good luck, see you tomorrow
+18:36 < Callum__> Alright, cya tomorrow them
+--- Day changed Mon Aug 12 2013
git.ucc.asn.au / matches / MCTX3420.git / blobdiff
UCC git Repository :: git.ucc.asn.au