Meeting started by phillw at 23:24:47 UTC. The full logs are available at http://mootbot.libertus.co.uk/to-make-impossible-possible/2010/to-make-impossible-possible.2010-12-10-23.24.log.html .
LINK: http://www.youtube.com/watch?v=fQ59dXOo63o (phillw, 23:38:14)
ACTION: hajour sketch up the screen shots (phillw, 23:43:18)
LINK: https://wiki.ubuntu.com/SpeechRecognition (phillw, 23:43:54)
LINK: https://blueprints.launchpad.net/ubuntu/+spec/speech-recognition (AlanBell, 23:44:03)
LINK: https://wiki.ubuntu.com/SpeechRecognition/SpeechMaker (AlanBell, 23:44:27)
LINK: http://bloc.eurion.net/archives/2010/espeak-gui-0-2/comment-page-1 (phillw, 23:45:09)
Meeting ended at 23:46:41 UTC.
- hajour sketch up the screen shots
People present (lines said)
23:24:47 <phillw> #startmeeting
23:24:47 <meetingology> Useful Commands: #topic #action #link #idea #voters #vote #chair #action #agreed #help #info #endmeeting.
23:24:48 <AlanBell> phillw: err, no
23:24:48 <Siekacz> ubuntu's not dropping GNOME
23:24:55 <Siekacz> just making a defferent shell
23:24:59 <AlanBell> ubuntu isn't in any way dropping gnome
23:25:17 <phillw> i stand corrected
23:25:19 <AlanBell> unity is a gnome application
23:25:30 <AlanBell> or a GTK one at least
23:25:37 <UndiFineD> thanks AlanBell : i was going to say the same
23:25:58 <phillw> but as a lubuntu person, the application should be independent of the DE
23:26:34 <Ryu_Kurisu> phillw: Yes, should be DE independent, however does anybody know how to do that?
23:26:42 <UndiFineD> so it would have a separate gtk / qt / ... interface
23:27:00 <AlanBell> really, the GUI is the least of the problems
23:27:08 <UndiFineD> the main app is just an api to plug into
23:27:09 <phillw> Ryu_Kurisu: python seems fairly happy.
23:27:16 <Ryu_Kurisu> (And even thou I'm going to switch to Arch (with LXDE), I would like to be DE independent)
23:27:45 <phillw> AlanBell: indeed, it is more about the coding of it, making it "look pretty" is on the wish list for the moment.
23:27:51 <Ryu_Kurisu> phillw: about what? DE independency?
23:28:00 <AlanBell> python doesn't dictate the widget toolkit to use or make it independent
23:28:10 <AlanBell> there is py-gtk and py-qt
23:28:41 <UndiFineD> the main app is just an api to plug into
23:28:44 <AlanBell> but really it doesn't take much to knock together a nice UI in quickly with glade
23:28:44 <Ryu_Kurisu> I know there is Vala gtk, don't know (nor really care :P) about Qt
23:29:01 <Ryu_Kurisu> Glade?
23:29:14 <phillw> AlanBell: is it difficult to get a python script compiled with both systems?
23:29:17 <AlanBell> glade is the UI designer for gtk
23:29:28 <Ryu_Kurisu> Alright, didn't know that...
23:29:38 <Ryu_Kurisu> (I've got ValaIDE atm)
23:30:17 <AlanBell> python is more interpreted than compiled, I guess it could work with both systems, but you would have to be careful with dependencies
23:30:49 <AlanBell> you don't want it to haul in all the dependencies of KDE when running on gnome just because it could run on KDE if you wanted it to
23:30:59 <Ryu_Kurisu> Correct me if I'm wrong, but with dbus you can do a lot, if not all the function we want?
23:31:00 <AlanBell> however this is largely irrelevant
23:31:26 <AlanBell> it probably doesn't need a GUI at all
23:31:28 <Ryu_Kurisu> UI with this is a secondary concern I agree
23:31:53 <phillw> As we seem to have more python people on board than C/C++ people do you think it is possible with python, possibly referring to C/C++ libraries? It's been too long since I did programming, so I'm not even sure what is possible.
23:32:01 <AlanBell> if the approach is to start from scratch then the project won't go anywhere at all
23:32:36 <phillw> AlanBell: +1
23:32:43 <UndiFineD> +1
23:32:54 <Ryu_Kurisu> But I was referring to when talking about DE Independency...I was talking about the starting and stopping of app, how things work in KDE or GNOME is just slightly different
23:33:32 <AlanBell> doing speech synthesis and recognition is beyond my capabilities, I simply don't know where to start
23:34:14 <AlanBell> in fact I do, I would go back to uni for a 3 year course in linguistics and computation
23:34:23 <Ryu_Kurisu> It's like I said before, we need to select a engine which we can use...at the same time would could define use-cases
23:34:32 <phillw> I know a guy who has experience of speech synthesis, but speech recognition seems to be the Holy Grail.
23:34:33 <AlanBell> then apply for a research position at the edinburgh speech centre
23:34:50 <AlanBell> then after a year or so at that I might be able to start doing useful work on it
23:35:27 <AlanBell> synthesis is easier than recognition, certainly
23:35:54 <AlanBell> and really rudimentary approach isn't too hard (dictionary of sound files, word matching and playing back the words)
23:36:32 <AlanBell> but that will sound rubbish, and won't have a sufficient vocabulary at all
23:36:34 <Ryu_Kurisu> AlanBell: I agree that we don't need to start from scratch, there are already some speech recognition engines out there which we could use...
23:36:43 <AlanBell> you need to do phoneme based stuff at the minimum
23:37:19 <AlanBell> the CMU sphynx library looks a good basis for recognition
23:37:21 <Ryu_Kurisu> I remember a kinect video which uses voice control to ask a computer things...just a sec
23:37:22 <phillw> As we are asking that it be able to have a book dictated to it, is should be a little easier?
23:37:47 <Ryu_Kurisu> First link: http://www.youtube.com/watch?v=fQ59dXOo63o
23:37:58 <AlanBell> if you have a domain specific vocabulary things can be simplified a little
23:38:14 <phillw> #link http://www.youtube.com/watch?v=fQ59dXOo63o
23:38:45 <Ryu_Kurisu> phillw: Thanks, found it already (or was that for the meeting?)
23:39:04 <phillw> Ryu_Kurisu: that was for meeting notes
23:39:12 <Ryu_Kurisu> Alright
23:41:30 <Ryu_Kurisu> I know from uni that it's usefull to have use-cases...not only can you direct your programming better, you also have got some good requirements you need to perform...
23:42:10 <AlanBell> it would be a good idea to update one of the existing specifications
23:42:21 <AlanBell> document the requirements
23:42:30 <phillw> Can you good people have a dig around the WWW and projects that we can link to. I will be in contact with the french guy once this meeting has finished to see if he will also come on board for the project.
23:42:34 <Ryu_Kurisu> We've got specifications already?
23:42:38 <AlanBell> sketch out any user interfaces using pencil or whatever
23:42:51 <AlanBell> Ryu_Kurisu: sure, everyone comes up with this idea
23:43:16 <AlanBell> and research the libraries that are already available and do the really really hard parts
23:43:18 <phillw> #action hajour sketch up the screen shots 23:43:18 * meetingology hajour sketch up the screen shots
23:43:21 <Ryu_Kurisu> Yeah, but that doesn't mean "we" have got specifications...
23:43:54 <phillw> #link https://wiki.ubuntu.com/SpeechRecognition
23:44:03 <AlanBell> #link https://blueprints.launchpad.net/ubuntu/+spec/speech-recognition
23:44:27 <AlanBell> #link https://wiki.ubuntu.com/SpeechRecognition/SpeechMaker
23:45:09 <phillw> #link http://bloc.eurion.net/archives/2010/espeak-gui-0-2/comment-page-1
23:45:51 <Ryu_Kurisu> Sorry guys/girls, I'm exhausted (live in Holland like hajour and is quarter to one here) so I'm off...
23:46:08 <Ryu_Kurisu> See ya'll tomorrow
23:46:20 <phillw> ;bye | Ryu_Kurisu
23:46:21 <DragonEyes> Ryu_Kurisu: take care and ensure the penguin does not catch you
23:46:31 <AlanBell> I am off too o/
23:46:41 <phillw> #endmeeting