18:07:24 <amber> #startmeeting
18:07:24 <tota11y> Meeting started Fri Feb 22 18:07:24 2013 CET.  The chair is amber. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:07:24 <tota11y> Useful Commands: #action #agreed #help #info #idea #link #topic.
18:07:37 <amber> so first off is status
18:07:41 <amber> #topic Status
18:07:57 <ghislop> I'm also guessing that this might be a shorter meeting since you met on Wednesday...
18:08:06 <Dark_Rose> Still reading through the meeting notes, starting to make a little more sense but not much more.
18:08:36 <amber> what spurred going through the meeting notes? just curious.
18:09:02 <Stoney> hi all
18:09:14 <amber> ghislop: I think we have some strageties to go over
18:09:23 <Stoney> sorry, heidi and i were in a meeting
18:09:28 <amber> welcome :)
18:09:29 <ghislop> hi Stoney... and Heidi.... we just started the meeting...
18:09:34 <heidi> :-) Yes, good.
18:09:37 <Dark_Rose> Heidi told me to look to the community and start reading through the notes and such left behind. Besides the actual code documentation I have not found much else online, so the next best thing was seeing how they progressed.
18:09:43 <ghislop> amber got volunteered by me to MC...
18:10:01 <Dark_Rose> Since you guys are looking at all the code, I figured I can look elsewhere for now.
18:10:28 <heidi> Sure, great!
18:10:29 <amber> ok. I'm wondering if we could start looking at gui stuff while everyone else is focused on opencv
18:11:05 <amber> its something I haven't started looking into yet, but is prolly the second biggest part of this program
18:11:18 <amber> just a thought
18:11:29 <amber> john?
18:11:55 <amber> do you have an update?
18:12:01 <Stoney> what do you mean by "look at" ?
18:12:27 <Stoney> or "looking into"?
18:12:30 <amber> start looking at what they are using for the gui, I know that opencv has their own gui built in but
18:12:33 <john_> Not really -- still reading the 1000+ page Learning Python book
18:12:38 <amber> im not sure if its being used
18:12:38 <Stoney> ok
18:13:16 <john_> Since it's using PyGTK, I'm thinking they're using the GTK hooks for GUI stuff?
18:14:03 <amber> I think so, but we wanted to upgrade it to gtk3 right? So we might want to start looking to see if thats possible and how it will effect the opencv upgrade
18:14:11 <amber> or vice versa
18:15:01 <john_> PyGTK isn't compatible with GTK3?
18:15:26 <amber> so as of now, I have been doing some documentation on the mousetrap project, primarily focusing on the framework and how it operates. I think there is a lot of room for consolidation or rewrites. Also just as an fyi I included a definitions list for opencv if anyone gets stuck on something
18:15:49 <amber> i dont know john
18:17:16 <john_> I'd be surprised if it is not (unless PyGTK was deprecated in favor of something else)
18:17:21 <john_> Python is popular those days...
18:18:26 <amber> PyGTK will be phased out with the transition to GTK+ version 3. Instead, GObject introspection will be used to generate bindings for Python and other languages on the fly. This is expected to eliminate the delay between GTK updates and corresponding language binding updates, as well as reduce maintenance burden on the developers.
18:18:56 <amber> says wikipedia
18:18:59 <john_> Weird -- Mousetrap currently uses BOTH PyGTK and GObject…?
18:19:03 <amber> im looking at their source now..
18:19:54 <heidi> Amber, did you have an idea of who particular should be doing the "looking"?
18:20:24 <amber> not really
18:20:28 <amber> i could start
18:21:28 <ghislop> amber: I'm open to suggestion on this.  If you'd like to poke in that direction, feel free.
18:21:59 <amber> I think its definitely the next step
18:22:10 <amber> I can put together some documentation on it
18:22:20 <amber> for when we start the transition
18:22:30 <john_> Any way we can just stick with a simple command line application with a default OpenCV-generated window for the webcam input for now, until we get the underlying stuff working?
18:23:13 <amber> I'm not changing the gui, just getting to know how it works
18:23:48 <amber> but yes, we will definitely keep is simple john
18:23:49 <john_> We'll have to change the code if we're punting PyGTK...
18:24:41 <amber> but not before we get opencv working
18:25:25 <ghislop> amber:  how does looking at the GUI fit with the issue of framework for opencv?  that is, should we keep you on the opencv stuff until we decide where we're headed?
18:26:33 <amber> the gui is an easy fix, but the actual face recognition i think is main priority
18:26:33 <amber> is*
18:26:33 <amber> can we all agree on that?
18:26:33 <amber> opencv first, then gui
18:26:58 <heidi> Yes.
18:27:09 <ghislop> +1
18:27:16 <john_> +2
18:27:21 <heidi> Let me ask a question. Is there no architecture or class diagram for Mousetrap?
18:27:29 <heidi> I haven't seen one, has anyone else?
18:27:32 <amber> its autogenerated from the code
18:27:42 <amber> doesnt explain how they fit together really
18:27:55 <amber> or how we would go about changing it
18:28:00 <heidi> What is autogenerated from the code?
18:28:04 <heidi> Is there a link?
18:28:33 <amber> I've been working on a better version here: http://www.xcitegroup.org/foss2serve/index.php/MouseTrap_Dev_Help
18:28:53 <amber> but gnome has one available:  http://gnome-mousetrap.sourcearchive.com/documentation/0.3plus-psvn17-2ubuntu1/main.html
18:29:11 <heidi> Ah, right. I was looking for a diagram.
18:29:20 <amber> there are a couple diagrams
18:29:27 <amber> that show the backend working together
18:29:42 <amber> http://gnome-mousetrap.sourcearchive.com/documentation/0.4-2/classmousetrap_1_1ocvfw_1_1__ocv_1_1OcvfwBase__inherit__graph.png
18:29:44 * heidi looking
18:31:26 <heidi> So it doesn't look like we have one, high-level picture with boxes and lines that represents the project?
18:31:32 <Stoney> Very incomplete
18:31:45 * ghislop looking too... amber: might want to record these as #links
18:31:58 <amber> they are all on the dev help site
18:32:11 <amber> #http://www.xcitegroup.org/foss2serve/index.php/MouseTrap_Dev_Help
18:32:15 <amber> #link http://www.xcitegroup.org/foss2serve/index.php/MouseTrap_Dev_Help
18:33:10 <amber> I could put together a class hierarchy, but can we upload pics to the wiki?
18:33:16 <heidi> :-)
18:33:30 <heidi> Ummm, ghislop is working on this.
18:33:58 <amber> its still in progress?
18:34:29 <ghislop> picture upload is off by default... it's a config change but I just haven't gotten to it.
18:34:43 <Stoney> might look for a UML generator for Python... here is a quick first effort (to look for one) ...
18:34:47 <Stoney> #link http://stackoverflow.com/questions/260165/whats-the-best-way-to-generate-a-uml-diagram-from-python-source-code
18:35:00 <amber> thanks
18:35:29 <amber> should we change to Next Steps?
18:35:38 <Stoney> actually, here is what that article links to...
18:35:41 <Stoney> #link http://epydoc.sourceforge.net/
18:35:41 <heidi> So yes,I think that the OpenCV issue should be resolved before we get into GUI stuff. And having documentation is going to helpeveryone.
18:35:42 <john_> 'closed as not constructive'
18:35:45 <john_> heh
18:36:11 <Stoney> yeah, that's the stackoverflow overlords for you :)
18:36:41 <amber> heidi, so I'm still going forward with the docs then for th gui?
18:36:51 <Dark_Rose> once we get close to finishing opencv, i'll start to look into the GUI stuff as well, I just want to see if I can find anything useful in the meeting notes, especially more recent ones that talk about cv2
18:37:02 <amber> #topic Next Steps
18:37:03 <heidi> Amber, I think that makes sense.
18:37:39 <heidi> I don't have the big picture of the MouseTrap architecture and without that, I think that it is difficult to see how changes in one area might impact others.
18:37:52 <amber> ok
18:38:20 <amber> so I have mousetrap program documentation, ocvfw documentation, diagram and gui documentation
18:38:28 <ghislop> It looks like epydoc requires comments in the code with some markup embedded in them... might not be our answer...  someone can look more closely though
18:38:55 <amber> i think its the lack of comments that makes their autogen docs so scarce
18:39:09 <amber> so it might be a manual effort
18:39:32 <Stoney> ghislop ... but does it need that for a diagram?
18:40:07 <Stoney> i think comments would be for a hyperlinked API documentation, but not necessarily for the diagram
18:40:18 <ghislop> Stoney:  not sure... someone will need to look further.  I'm a 30 second expert here :-)
18:40:23 <Stoney> yup yup
18:40:26 <Stoney> me too :)
18:40:29 <amber> ill check it out
18:41:02 <amber> and dark_rose, if you need some clarifications on verbiage, I created a definition list for opencv
18:41:17 <Dark_Rose> ok, and that is all on the dev page?
18:41:21 <amber> describes how the go about creating haars and haar detection
18:41:22 <amber> yes
18:41:43 <heidi> Yes,, thank you for creating that Amber. Helpful!
18:41:58 <Dark_Rose> alright, thanks!
18:42:03 <amber> :)
18:42:24 <amber> john, any next steps?
18:43:13 <john_> No clue
18:44:06 <amber> I suggest using the online python resources instead of a book
18:44:12 <amber> the tutorials are very helpful
18:44:13 <amber> http://www.python.org/
18:44:18 <amber> (im sure you've been there)
18:44:50 <amber> and also looking into the ocvfw, looking at the documentation I put together and looking for points to consolidate/remove
18:45:09 <amber> if you see things missing in the docs, feel free to add
18:45:37 <john_> The book came highly recommended. Begun out good, then got boggled by details. I'll try the online tutorials -- I prefer to learn the basics, then learn stuff as I need them, anyway
18:46:03 <john_> ocvfw = all the .py files in the ocvfw folder, right?
18:46:25 <Stoney> (it looks like epydoc can be install through yum, apt, etc.)
18:46:28 <amber> yes. ocvfw is the folder that contains everything for the OpenCV FrameWork
18:46:44 <amber> o nice, I can mess with epydoc after this then
18:47:41 <amber> so just to bring up one last topic that we touched on last meeting
18:47:55 <Dark_Rose> got to go to class, let me know if i miss anything! bye everyone
18:47:59 <amber> is about the framework. Is the goal to rewrite, or revamp?
18:48:03 <amber> bye!
18:48:19 <john_> I remember you mentioning we might delete that folder entirely
18:48:24 <ghislop> do we know enough to answer that question?
18:48:30 <amber> getting there
18:48:38 <Stoney> i think  we need to understand the current framework
18:48:46 <amber> I know we will need some type of wrapper-ish thing around opencv
18:48:48 <Stoney> maybe epydoc will help
18:49:06 <amber> but the current one is redundant in many places
18:49:39 <heidi> Right.
18:49:44 <amber> ok, ill look into that then
18:49:55 <heidi> I think we should come up with our best guess of what we should do and then run it by the community.
18:49:57 <Stoney> so ocvfw is a lost cause?
18:50:00 <john_> amber, I had to leave early on Wednesday, what did I miss?
18:50:13 <Stoney> heidi, that's an idea
18:50:13 <amber> just this i think john.
18:50:19 <ghislop> so that sounds like we need to roughly estimate the change to fix the framework... and based on that decide whether revision or replacement is the easier approach
18:50:21 <amber> yes i like that heidi
18:50:30 <heidi> :-)
18:50:33 <amber> right ghislop
18:50:49 <amber> so well wait for the docs, and then proceed
18:50:49 <heidi> The community may also have insight into which is a better approach and how to approach the problem.
18:50:53 <heidi> :-)
18:51:03 <ghislop> amber: I think you've got the ball on the framework issue.
18:51:16 <amber> ok.
18:51:22 <Stoney> getting an estimate of the changes to the current framework is probably easier than estimating what a rewrite will look like
18:51:23 <amber> so ill make some action items
18:51:28 <ghislop> why don't you work on the assessment and then share results with everyone?
18:51:38 <amber> ok
18:52:03 <amber> #action Dark_Rose will add documentation about cv2 based on opencv meeting notes
18:52:09 <ghislop> Stoney: agreed.  but if the revision looks messy enough, I think that's the argument for repacement
18:52:13 <amber> #action Amber will create a diagram of mousetrap
18:52:28 <Stoney> good point ghislop
18:52:47 <amber> I think they might have a structure we could use
18:53:03 <amber> and some of the nitty-gritty stuff in opencv is something I don't want to touch
18:53:51 <amber> #action Amber will create documentation on the mousetrap program, gui and ocvfw
18:54:07 <amber> #action john will continue to learn python and ocvfw
18:54:13 <amber> we are all ok with these actions?
18:54:17 <amber> am i missing something?
18:54:29 <ghislop> sounds good to me.
18:54:40 <john_> What exactly would you like me to do with ocvfw?
18:54:47 <amber> learn how it operates
18:55:02 <john_> Okay I'll try my best :)
18:55:03 <amber> because the next step is coding with it
18:55:27 <john_> Wait, are we going to use it or are we going to remove/replace it?
18:55:34 <ghislop> I'm assuming that "learn how it operates" is from an external perspective?
18:55:37 <amber> thats what I will find out
18:56:01 <amber> what do you mean by external?
18:56:16 <john_> I'm assuming I'll take a look at the .py files and basically assemble the functions and what they do?
18:56:35 <ghislop> enough to understand how it's used by mousetrap, not a focus on the internals of the code?
18:57:08 <john_> That'll be a complicated task, since I don't know what .py files use what functions...
18:57:09 <amber> I think in order to understand how its used, you will have to look at the internals
18:57:27 <amber> I have it documented john, you can just look over what I have on the dev page
18:57:47 <john_> Okay
18:58:04 <amber> I am currently working on how the .py files operate, I think you can just take a look at my docs and that should be fine
18:58:26 <amber> does that make sense?
18:58:34 <john_> I think so
18:58:45 <amber> so you wont be looking at actual code, just the docs
18:59:15 <amber> so we all know what we're doing ?
18:59:36 <john_> I think I'll look at actual code -- that should help my learning along, too. I also want to see what the functions the functions use (if that makes sense)
19:00:08 <ghislop> OK
19:00:10 <amber> I do have the methods on there too, you could actually do a side-by-side
19:00:17 <john_> lol
19:00:23 <john_> you've done all the work already :p
19:00:33 <amber> its not finished yet :)
19:00:40 <amber> mostly done for ocvfw though
19:01:21 <amber> so I think thats it. anything else?
19:02:02 <amber> awesomes.
19:02:04 <amber> #endmeeting