Committee/Meetings/2008 11 11
From AnnodexFoundation
No official minutes were taken at this meeting, but here is the irc log.
Nov 11 14:04:12 <nessy> hi everybody
Nov 11 14:04:23 <nessy> it's time for the public annodex committee meeting
Nov 11 14:04:29 <nessy> who's here to take part?
Nov 11 14:04:51 <nessy> Agenda:
Nov 11 14:04:59 <nessy> 1. Update on projects
Nov 11 14:05:05 <nessy> 2. Update on standards
Nov 11 14:05:10 <nessy> 3. Update on collaborations
Nov 11 14:05:17 <nessy> 4. Update on FOMS
Nov 11 14:05:17 <nessy> 5. Time/Location for AGM
Nov 11 14:05:17 <nessy> 6. Other business
Nov 11 14:06:24 <kfish> 7. The Larch
Nov 11 14:06:50 <nessy> purserj, MikeS, shans - any of you here?
Nov 11 14:06:52 <nessy> hi kfish!!
Nov 11 14:07:01 * mdale :No such nick/channel
Nov 11 14:07:20 <nessy> seems mdale and dennisc can't make it
Nov 11 14:07:48 <nessy> anyone non-committee here to participate? :-)
Nov 11 14:08:10 <doublec> i'm here
Nov 11 14:08:37 <nessy> heya!
Nov 11 14:08:48 <nessy> awesome - we can have a chat amongst us :)
Nov 11 14:09:01 <nessy> let's get started so I can get some lunch
Nov 11 14:09:08 <nessy> 1. Update on projects
Nov 11 14:09:34 <nessy> wiking are you here, too?
Nov 11 14:09:54 <nessy> liboggplay seems to get towards a maintainer
Nov 11 14:10:11 <nessy> I'm very happy that wiking is getting so active
Nov 11 14:10:24 <nessy> doublec - wanna give an update on firefox etc
Nov 11 14:10:38 <nessy> ?
Nov 11 14:10:41 <doublec> yep
Nov 11 14:10:53 <doublec> Support for Theora and Vorbis playback is steaming ahead.
Nov 11 14:11:06 <doublec> FF beta 2 is coming out soon
Nov 11 14:11:17 <doublec> there's been a lot of discussion on cross domain restriction of video
Nov 11 14:11:20 <doublec> but nothing decided yet
Nov 11 14:11:25 <doublec> a thread on WHATWG list has started
Nov 11 14:11:43 * Magical_Musings (n=pinkduck@cpe-76-184-143-198.tx.res.rr.com) has joined #annodex
Nov 11 14:11:45 <doublec> That's about it I think
Nov 11 14:11:54 * Magical_Musings (n=pinkduck@cpe-76-184-143-198.tx.res.rr.com) has left #annodex
Nov 11 14:12:06 <nessy> wiking fixed the yuv2rgb bug, right?
Nov 11 14:12:15 <purserj> nessy: I'm here but am about to leave sorry
Nov 11 14:12:29 <nessy> purserj: no worries
Nov 11 14:12:37 <nessy> I might send an irclog
Nov 11 14:12:55 <nessy> and a whole list of new bugs has been discovered
Nov 11 14:13:02 <doublec> nessy, it's fixed for the non-optimized case
Nov 11 14:13:15 <doublec> but iirc there may be a problem for the optimized solution
Nov 11 14:13:21 <doublec> but I don't use that, so it doesn't affect me
Nov 11 14:13:28 <nessy> cool
Nov 11 14:13:39 <nessy> I do think wiking's your man now
Nov 11 14:13:41 <doublec> fuzz testing has started on all the third party libraries too
Nov 11 14:13:49 <doublec> so there may be bugs related to that popping up
Nov 11 14:14:03 * denisc (n=dcrowdy@knakazawa.mdlang.mq.edu.au) has joined #annodex
Nov 11 14:14:09 <nessy> hi denisc!
Nov 11 14:14:17 <denisc> Hi! Sorry I'm late :)
Nov 11 14:14:39 <nessy> third party libraries - such as libtheora, libvorbis, libogg and liboggz
Nov 11 14:14:40 <nessy> ?
Nov 11 14:15:03 <doublec> libtheora, libvorbis, libogg, liboggz, libfishsound, liboggplay
Nov 11 14:15:11 <nessy> cool
Nov 11 14:15:30 <nessy> is it important for you that regular releases are made of liboggplay or is svn enough for you?
Nov 11 14:15:38 <doublec> svn is fine
Nov 11 14:15:50 <nessy> cool
Nov 11 14:16:23 <nessy> kfish: over to you - liboggz, oggz-tools etc
Nov 11 14:16:34 <kfish> right
Nov 11 14:16:37 <MikeS> here, sorry
Nov 11 14:16:59 <nessy> hi MikeS - awesome! now we have 4 committee members!
Nov 11 14:17:04 <kfish> random maintenance i guess, mostly on oggz-chop and recently on libannodex also
Nov 11 14:17:12 <nessy> (5 with purserj)
Nov 11 14:17:38 <nessy> yeah - I was wondering about that
Nov 11 14:17:48 <kfish> i guess it's about time for a liboggz release
Nov 11 14:18:02 <nessy> I thought you were going to abandon libannodex
Nov 11 14:18:12 <nessy> why the fixes?
Nov 11 14:18:25 <kfish> i was hacking up a test browser recently for playing with opensearch
Nov 11 14:18:59 <kfish> which i've implemented before in cmmlwiki, so i was testing against that
Nov 11 14:19:33 <kfish> and the presentation time and startgranules weren't getting written into skeleton, so i fixed that in libannodex
Nov 11 14:20:00 <nessy> how are the plans to combine oggz-chop, oggz-rip, oggz-merge into a media server?
Nov 11 14:20:02 <kfish> the code's not so bad, once i throw half of it away and rewrite the other half ;-)
Nov 11 14:20:14 <nessy> lol
Nov 11 14:20:32 <kfish> so, basically that will involve putting some of those tool functions into liboggz
Nov 11 14:20:42 <kfish> so that they are easier to combine (like they are in hogg)
Nov 11 14:21:20 <kfish> i'm writing up some notes about ogg chopping also of course (for lca etc.)
Nov 11 14:21:31 <nessy> using oggz-chop?
Nov 11 14:21:44 <kfish> which?
Nov 11 14:22:12 <nessy> well, I guess I am a bit confused as to what we should recommend people to use for chopping on their servers
Nov 11 14:22:16 <kfish> well, just what the process is and how oggz-chop does it, yeah
Nov 11 14:22:21 <kfish> oggz-chop!
Nov 11 14:22:30 <kfish> is the fastest
Nov 11 14:22:51 <nessy> so to get an all-competent media server, what is the plan?
Nov 11 14:22:59 * Magical_Musings (n=pinkduck@cpe-76-184-143-198.tx.res.rr.com) has joined #annodex
Nov 11 14:23:18 <kfish> well, that depends what you mean by "all-competent"
Nov 11 14:23:23 * Magical_Musings (n=pinkduck@cpe-76-184-143-198.tx.res.rr.com) has left #annodex
Nov 11 14:23:36 <nessy> the track extraction and recomposition using ROE etc
Nov 11 14:23:41 <kfish> right
Nov 11 14:24:06 <nessy> is it best to write that with a combination of oggz-chop, oggz-rip, oggz-merge?
Nov 11 14:24:32 <nessy> and thus we totally throw away mod_annodex?
Nov 11 14:24:42 <kfish> it could be done with those and a bunch of scripts, but that would be sucky
Nov 11 14:25:24 <kfish> well, if our aim is to support ROE server-side (for constructing a multiplexed stream), then mod_annodex isn't really relevant anyway
Nov 11 14:25:50 * nessy nods
Nov 11 14:25:59 <nessy> yeah, I guess new code is needed anyway
Nov 11 14:26:18 <doublec> will oggz-chop support byte range requests?
Nov 11 14:26:28 <kfish> probably what i'll do is prototype it with hogg, which will be cleaner, then re-implement in C in liboggz
Nov 11 14:26:38 <kfish> doublec, yes, that's definitely the plan
Nov 11 14:26:43 <doublec> cool!
Nov 11 14:27:00 <nessy> maybe we should adopt the new media fragments draft as soon as it's out and develop towards that
Nov 11 14:27:41 <kfish> nessy, that relates to URI syntax?
Nov 11 14:27:41 <nessy> which is very similar to our original plans for time range requests
Nov 11 14:27:44 <nessy> yes
Nov 11 14:28:00 <nessy> and to web proxy support with byte ranges and all
Nov 11 14:28:21 * kfish wonders where this draft is
Nov 11 14:28:21 <nessy> doublec: are there plans to support the byte ranges in firefox?
Nov 11 14:28:35 <doublec> nessy, firefox already does
Nov 11 14:28:47 <nessy> kfish: it's developing here http://www.w3.org/2008/WebVideo/Fragments/wiki/Main_Page
Nov 11 14:28:50 <doublec> seeking works using byte range requests
Nov 11 14:28:53 <nessy> (see my blog post from yesterday)
Nov 11 14:28:58 <kfish> ok!
Nov 11 14:29:08 <nessy> doublec: that's awesome!
Nov 11 14:29:20 <nessy> how do you map time to bytes in the browser?
Nov 11 14:29:44 <nessy> i.e. if somebody hit's a time offset for playback, how do you map that to bytes?
Nov 11 14:29:56 <nessy> a rough guess based on file lenght and bitrate?
Nov 11 14:30:28 <nessy> or are you talking about the case where you do it all client side?
Nov 11 14:30:43 <doublec> nessy, I use liboggz to do the seeking
Nov 11 14:31:01 <nessy> ah, so it's only for the case where you have already received all the data
Nov 11 14:31:02 <doublec> but the i/o is done with byte range requests
Nov 11 14:31:06 <doublec> instead of file i/o
Nov 11 14:31:15 <doublec> unless it's a local file on the disk of course
Nov 11 14:31:22 <doublec> nessy, no, it doesn't need all the data
Nov 11 14:31:37 <doublec> so if a seek request is made by the user for time position 30s
Nov 11 14:31:41 <doublec> I ask libogg play for this
Nov 11 14:31:47 <doublec> it makes a guess at a byte position
Nov 11 14:31:53 <doublec> I do a http byte range request and return that data
Nov 11 14:31:59 <doublec> based on that liboggz does more guesses, etc
Nov 11 14:32:17 <nessy> I see - it is thus based on the guesswork
Nov 11 14:32:25 <doublec> When the video is first loaded I actually seek to the end via byte range requests to get the duration too
Nov 11 14:32:38 <doublec> yes
Nov 11 14:32:39 <nessy> via byte range requests - wow!
Nov 11 14:32:47 <nessy> do you have that documented somewhere?
Nov 11 14:32:48 <kfish> do you have those byte-range requests hooked up as OggzIO callbacks?
Nov 11 14:32:58 <nessy> that would be really important for me to share with the W3C guys
Nov 11 14:33:48 <doublec> kfish, the OggzIO callbacks are hooked up to liboggplay
Nov 11 14:33:56 <doublec> which calls it's own 'readers'
Nov 11 14:34:09 <doublec> We have a 'reader' implementation that uses the firefox network api
Nov 11 14:34:15 <kfish> ah cool, got it
Nov 11 14:34:34 <kfish> that's pretty neat
Nov 11 14:34:57 <doublec> which basically uses this class http://hg.mozilla.org/mozilla-central/file/5dfdad637696/content/media/video/src/nsMediaStream.cpp
Nov 11 14:35:24 <doublec> which optimizes for local file usage vs http usage
Nov 11 14:35:36 <doublec> the plan is to detect if the file is in the browser cache and use fast file seeks in that case
Nov 11 14:35:40 <doublec> but it's not implemented yet
Nov 11 14:35:58 <nessy> oh I see - you're doing the seeking over the network with built-in guesswork
Nov 11 14:36:06 <doublec> yes
Nov 11 14:36:08 <kfish> heh, sounds like fun
Nov 11 14:36:24 <nessy> doesn't that - like - create a lot of network requests for one seek?
Nov 11 14:36:50 <MikeS> only if you want accurate seeking! Nobody cares about accurate seeking.
Nov 11 14:37:50 <doublec> nessy, yes, it can do a few
Nov 11 14:37:52 <nessy> any stats on how many network connections you're getting per seek?
Nov 11 14:38:17 <doublec> if the guesses are good, less than 10
Nov 11 14:38:24 <nessy> (cause: I spoke to TBL when I was in France and he is really concerned about Web proxies doing more than one roundtrip to get the fragment data)
Nov 11 14:38:26 <kfish> is it re-using a persistent connection?
Nov 11 14:38:30 <doublec> in practice it works fairly fast
Nov 11 14:38:56 <doublec> it should do
Nov 11 14:39:01 <doublec> there are issues that need to be resolved
Nov 11 14:39:12 <doublec> like if a malicious web server serves a different file on the byte range request
Nov 11 14:39:19 <doublec> maybe it can do some sort of exploit
Nov 11 14:39:26 <nessy> doublec: our suggestion is rather to tell the server what time range you want and it returns you which byte range that maps to and the byte data
Nov 11 14:39:29 <doublec> and there are other optimizations we can doo
Nov 11 14:39:53 <doublec> nessy, I had planned to detect if it was served by liboggz or libannodex and use the server side seek support
Nov 11 14:39:59 <nessy> cool
Nov 11 14:40:00 <doublec> but haven't yet got around to it
Nov 11 14:40:05 <nessy> no worries :)
Nov 11 14:40:14 <doublec> and I need to think how that interacts with parts of the whatwg api
Nov 11 14:40:22 <kfish> even so, for cache-friendliness we'll do the byte-range translation for the data section
Nov 11 14:40:23 <nessy> that would be an awesome thing to implement - and it would need support from the server, too
Nov 11 14:40:58 <kfish> and it sounds like i'll need to discuss that with the fragments wg ...
Nov 11 14:41:39 <nessy> kfish: there were massive discussion on the mailing list - see http://lists.w3.org/Archives/Public/public-media-fragment/
Nov 11 14:42:16 <kfish> ok i'll check it, i thought that wg was just dealing with the uri scheme, not the transport
Nov 11 14:42:21 <nessy> doublec: you do know that there is a spec in the annodex I-D about how to do the byte-range request in a way that enables proxies to cache them?
Nov 11 14:42:43 <nessy> kfish: it's about getting the URI mechanism rolled out so naturally it involves everything
Nov 11 14:42:50 <nessy> including the selection of tracks!
Nov 11 14:42:55 <doublec> nessy, is there? Do you have a link?
Nov 11 14:43:30 <nessy> doublec: section 6.2 on http://annodex.net/TR/draft-pfeiffer-temporal-fragments-03.html
Nov 11 14:43:45 <doublec> thanks!
Nov 11 14:44:01 <nessy> in fact, one proposal to the media fragments working group is exactly the mechanism specified there
Nov 11 14:44:03 <doublec> I've tested the seeking with wikimedia, and other sites, and it works well
Nov 11 14:44:18 <nessy> the other one is to empower web proxies to support time range requests directly
Nov 11 14:44:24 <nessy> awesome!
Nov 11 14:44:34 <nessy> do you have the procedure documented somewhere?
Nov 11 14:44:45 <doublec> to seek you just need to set the currentTime attribute of the video
Nov 11 14:44:53 <nessy> I'd like to point that out to all the ppl who scream 2 roundtrips are too slow :)
Nov 11 14:44:58 <doublec> document.getElementsByTagName("video")[0].currentTime=20
Nov 11 14:45:03 <doublec> the value is number of seconds
Nov 11 14:45:24 <nessy> coll - I think I can get it all out of the log now :)
Nov 11 14:45:25 <doublec> it'll have a user interface too (a scrubber bar) soon
Nov 11 14:45:33 <nessy> rockin'!
Nov 11 14:45:57 <nessy> ok, let's move on - I'm starving :)
Nov 11 14:46:04 <nessy> libcmml - I haven't done anything
Nov 11 14:46:11 <nessy> but I want to turn it into a codec like kate
Nov 11 14:46:28 <nessy> rather than doing cmml though, I'm working on a generic library for mapping text codecs into ogg
Nov 11 14:46:38 <nessy> first spec is here http://wiki.xiph.org/index.php/Text_Codecs_Ogg_Mapping
Nov 11 14:46:53 <nessy> still in the works - I'm discussing it lots with ogg.k.ogg.k
Nov 11 14:47:24 <nessy> the first thing that will work is srt in ogg
Nov 11 14:47:30 <nessy> will be really trivial with that spec
Nov 11 14:47:54 <kfish> awesome!
Nov 11 14:47:58 <nessy> all a "text codec" will need to do is define its header packets and its data packets
Nov 11 14:48:19 <nessy> then hand these to the library that will eventually evolve
Nov 11 14:48:36 <nessy> and it's encapsulated in a seekable way
Nov 11 14:48:43 <nessy> is lots of fun :)
Nov 11 14:49:22 <nessy> the harder part is to create a "API" or a "DOM model" for extraction that Firefox can use and know what to do with
Nov 11 14:49:44 <nessy> that will be the real fun :)
Nov 11 14:50:13 <nessy> doublec: I'd like to discuss that at some point with you
Nov 11 14:50:22 <nessy> what can be done, what should best be done etc
Nov 11 14:50:34 <doublec> nessy, sure, happy to discuss anytime
Nov 11 14:50:47 <nessy> cool!
Nov 11 14:51:14 <nessy> I'd also like to implement this library for Mozilla, but my time is limited and I have to achieve a whole lot before Christmas
Nov 11 14:51:27 <nessy> might need to chat to Aaron about it
Nov 11 14:52:02 <nessy> so, libcmml may be revived in very interesting ways :)
Nov 11 14:52:26 <nessy> as for ROE - I haven't done anything with it yet
Nov 11 14:52:43 <nessy> I think it will become interesting with the media fragments WG at some point
Nov 11 14:52:59 <nessy> and it may be rewritten as a SMIL spec by the SMIL experts :)
Nov 11 14:53:06 <nessy> but I don't mind - I just want that functionality
Nov 11 14:53:40 <nessy> ok - have we any other projects to discuss?
Nov 11 14:53:55 <nessy> ah yes - the new CSIRO IE plugin and DMTagger codebase
Nov 11 14:54:02 <nessy> they need maintainers
Nov 11 14:54:37 <nessy> I took a interview with purserj yesterday and he will publish the podcast next week - it contains an appeal for maintainers
Nov 11 14:55:38 <nessy> I also talked to Cristian Adam who is now the maintainer of oggcodecs
Nov 11 14:55:48 <nessy> but he has enough to do with just oggcodecs
Nov 11 14:56:19 <nessy> he will however package the stand-alone transcoder app out of dmtagger as its own application
Nov 11 14:56:41 <nessy> so that's pretty cool - to have a kind of ffmpeg2annodex for Windows
Nov 11 14:57:17 <nessy> anything else to report?
Nov 11 14:58:36 <nessy> 2. Update on standards
Nov 11 14:58:59 <nessy> accessibility:
Nov 11 14:59:09 <nessy> I'm continuing the work with Mozilla on that
Nov 11 14:59:21 <nessy> it's been really interesting to analyse all the use cases and come up with a list of requirements
Nov 11 14:59:30 <nessy> all information is here https://wiki.mozilla.org/Accessibility
Nov 11 14:59:45 <nessy> basically, it seems that kate would be perfect as the solution for basically all needs
Nov 11 15:00:00 <nessy> in particular since kate has included the whole layout and presentation idea
Nov 11 15:00:06 <nessy> which cmml has persistently ignored
Nov 11 15:00:26 <nessy> ups, link should be https://wiki.mozilla.org/Accessibility/Video_Accessibility
Nov 11 15:00:59 <nessy> there are a few things that I'd like to improve on the kate spec and some things that may be too complicated to ever implement
Nov 11 15:01:29 <nessy> so basically I am considering creating a mix of TimedText, Kate and CMML as a universal text codec to cover all needs
Nov 11 15:01:45 <nessy> but right now I don't even wanna go there, but just enable what exists to work
Nov 11 15:02:07 <nessy> ...
Nov 11 15:02:11 <nessy> anyone still listening?
Nov 11 15:02:48 <nessy> moving on ...
Nov 11 15:03:08 <nessy> at the W3C I am now in the media fragments, media annotations, and the timed text working groups
Nov 11 15:03:39 <nessy> media fragments - as discussed above - creates a URI mechanism including protocol mechanism to deliver media fragments over the Web
Nov 11 15:03:59 <nessy> recycles and develops further many ideas from our temporal URI specs
Nov 11 15:04:05 * fishkandy (n=conrad@meg.dl.kuis.kyoto-u.ac.jp) has joined #annodex
Nov 11 15:04:30 <nessy> media annotations - is about creating a metadata scheme for audio/video
Nov 11 15:04:41 <nessy> kindof a better vorbiscomment
Nov 11 15:04:45 <nessy> we'll see where that ends
Nov 11 15:04:54 <doublec> you joined all the hard w3c working groups I see
Nov 11 15:05:12 <nessy> timed text - has been revived and is about making a simpler timed text spec and getting implementations
Nov 11 15:05:29 <nessy> I'm also continuing to watch the HTML5 video/audio specs :)
Nov 11 15:05:38 <nessy> I have met Ian Hickson - rather cool dude
Nov 11 15:05:45 <doublec> yeah, he's a good guy
Nov 11 15:05:54 <nessy> and we had a chat about what what audio & video need
Nov 11 15:06:05 <nessy> will make some contributions as I learn more about stuff
Nov 11 15:06:25 <nessy> yeah - all the a/v stuff at W3C
Nov 11 15:06:49 <nessy> I gave a lightning talk about what HTML5 video is missing - made quite some impact at the meeting in France
Nov 11 15:07:05 <nessy> and I met many ppl from different areas - including accessibility
Nov 11 15:07:36 <nessy> also I heard about a new standards process at SMPTE
Nov 11 15:07:44 * kfish is now known as Guest42708
Nov 11 15:07:45 <nessy> they are also looking into timed text and annodations
Nov 11 15:08:01 <nessy> not sure I have the bandwidth to follow them, too
Nov 11 15:08:26 <nessy> enough on standards :)
Nov 11 15:08:29 <nessy> 3. Update on collaborations
Nov 11 15:08:47 * Guest42708 has quit (Read error: 110 (Connection timed out))
Nov 11 15:08:59 * fishkandy is now known as kfish
Nov 11 15:09:53 <nessy> just for everyone here who's not involved in the discussions - there are talks going on about Mozilla potentially sponsoring some ppl to do liboggplay/liboggz/theora type work
and help get the media support in FF3 more stable sooner
Nov 11 15:10:16 <nessy> it's really exciting to see so many new efforts in the space!
Nov 11 15:10:23 <kfish> worrd
Nov 11 15:10:37 <nessy> and above all this CSIRO has open sourced their remaining Annodex codebase
Nov 11 15:10:58 <nessy> AIDA seems to be lost from this planet though
Nov 11 15:11:19 <nessy> all I have are the old mobile phones from CISRO that have a binary version of AIDA running
Nov 11 15:11:29 <doublec> what was AIDA?
Nov 11 15:11:34 <kfish> it's almost like closed-source is a bad idea
Nov 11 15:11:36 <nessy> a mobile phone annodex player
Nov 11 15:11:46 <nessy> kfish: who would have thought ;-)
Nov 11 15:12:15 <nessy> I am still of the firm opinion that all code that is written at unis and research institutes should be open sourced
Nov 11 15:12:17 <kfish> doublec, a symbian player, though at least as a result of that our symbian patches went upstream
Nov 11 15:12:37 <nessy> yeah, I think the most important parts are in the libraries with the symbian patches
Nov 11 15:12:43 <doublec> nice
Nov 11 15:12:44 <nessy> I'm not too fussed about the application
Nov 11 15:13:02 <nessy> but we have two mobile phones that run symbian that we can give away to some developers at FOMS if we want
Nov 11 15:13:14 <nessy> (if anyone is really interested in those old phones still)
Nov 11 15:13:43 <kfish> yeah, they have to promise to maintain symbian support in all the xiph libraries, lol
Nov 11 15:13:49 <nessy> lol
Nov 11 15:14:12 <nessy> btw: CISRO also became a paying advisory board member of the Annodex association
Nov 11 15:14:16 <kfish> as a reward, they get a circa-2005 3G phone
Nov 11 15:14:27 <nessy> that's a really good outcome and a great relationship to CSIRO post their involvement with Annodex
Nov 11 15:15:14 <nessy> ok, moving on...
Nov 11 15:15:15 <nessy> 4. Update on FOMS
Nov 11 15:15:32 <nessy> http://www.foms-workshop.org/foms2009/
Nov 11 15:15:49 <nessy> anyone here should really think about coming
Nov 11 15:15:58 <nessy> we'll have all the open video codec developers there!
Nov 11 15:16:19 <nessy> dirac, theora & OMS (Sun's new one)
Nov 11 15:16:32 <nessy> also we'll hopefully have Mozilla and Opera present
Nov 11 15:16:41 <nessy> ffmpeg with possibly two people
Nov 11 15:16:47 <nessy> pitivi
Nov 11 15:16:59 <nessy> xiph/annodex of course
Nov 11 15:17:06 <nessy> and lots of other cool people and projects
Nov 11 15:17:48 <nessy> sponsorship is a bit slow this year - not surprising under these market conditions - but we'll still have a great workshop
Nov 11 15:18:16 <nessy> oh ... and there will be a really cool t-shirt !!!
Nov 11 15:18:32 <nessy> next...
Nov 11 15:18:34 <nessy> 5. Time/Location for AGM
Nov 11 15:19:01 <nessy> should we have the AGM this time around again at FOMS?
Nov 11 15:19:08 <nessy> we will soon have to prepare for election, btw
Nov 11 15:20:04 <nessy> we will need ppl to stand for committee next year - think about standing!
Nov 11 15:20:27 <kfish> same time (at foms) for the agm sounds good to me, and election online like previous years
Nov 11 15:20:34 <nessy> awesome
Nov 11 15:20:42 <nessy> can you set up that election site again?
Nov 11 15:20:46 <kfish> sure
Nov 11 15:20:47 <nessy> I think we need to start soon
Nov 11 15:20:57 <nessy> great!
Nov 11 15:21:31 <nessy> hmm... and I guess I need to send out an email for the elections
Nov 11 15:21:32 <kfish> it's about 10 weeks away, the election should probably open, 1 week before the agm?
Nov 11 15:21:38 <kfish> or 2 weeks?
Nov 11 15:21:43 <nessy> I think it was a month, no?
Nov 11 15:21:50 <kfish> a month for casting votes?
Nov 11 15:22:00 <nessy> hmm... I might be mistaken
Nov 11 15:22:22 <kfish> a month or more for taking nominations is good
Nov 11 15:22:44 <nessy> schedule from last year:
Nov 11 15:22:45 <nessy> > Dec 13 (6 weeks prior to the AGM): nominations open
Nov 11 15:22:45 <nessy> > Jan 3 (3 weeks prior to the AGM): nominations closed, voting open
Nov 11 15:22:45 <nessy> >
Nov 11 15:22:45 <nessy> > Jan 23 (1 day prior to the AGM): voting closed
Nov 11 15:22:45 <nessy> > Jan 24 (AGM): 2008 committee announced
Nov 11 15:23:11 <nessy> yeah, with Christmas around
Nov 11 15:23:29 <nessy> cool - I'll leave it with you, if that's ok :)
Nov 11 15:24:17 <nessy> 6. Other business
Nov 11 15:24:20 <kfish> cool
Nov 11 15:24:24 <nessy> thanks!
Nov 11 15:25:03 <nessy> anything else we should talk about?
Nov 11 15:25:23 <nessy> denisc: what've you been up to recently? :)
Nov 11 15:25:42 <denisc> Getting some more data into arkaiv
Nov 11 15:25:42 <doublec> there needs to be tools for generating content. Any thoughts about that?
Nov 11 15:25:52 <denisc> We have a person for a few months
Nov 11 15:26:07 <doublec> Once browsers support playback etc people will start asking how to best get their files into that format and onto servers
Nov 11 15:26:08 <denisc> Haven't abandoned arkaiv yet because it sort of works and is being used...
Nov 11 15:26:53 <denisc> Sorry doublec - double conversation! I have a tool for content generation that my assistant uses; could work it into somethig more sensible
Nov 11 15:26:59 <nessy> doublec: there's always ffmpeg2theora, right?
Nov 11 15:27:05 <nessy> there's a gui on the mac for it
Nov 11 15:27:18 <denisc> Can annotations be added as you go though?
Nov 11 15:27:30 <nessy> and that app on windows that Cristian is developing will also enable that
Nov 11 15:27:40 <doublec> ok, good that stuff is being done
Nov 11 15:27:47 <nessy> no, the annotation/caption etc stuff has not been implemented yet
Nov 11 15:28:26 <nessy> I'm hoping for some coding support somewhere :)
Nov 11 15:28:29 <kfish> denisc, more work on arkaiv is great, specialised server applications are goood for us :-)
Nov 11 15:28:56 <denisc> I guess so - but changing it to not rely on mod_annodex will be a challenge for me... Fun challenge though
Nov 11 15:29:04 <nessy> denisc: how cross-platform is archaiv?
Nov 11 15:29:22 <denisc> Depends on pylons which is CP I think?
Nov 11 15:29:27 <nessy> cool
Nov 11 15:29:37 <denisc> You mean as a user or as an app?
Nov 11 15:29:50 <nessy> maybe we should start a web site with applications and server software for theora
Nov 11 15:30:14 <denisc> I should actually do something so if there's anything simple that needs doing throw it my way
Nov 11 15:30:15 <nessy> or more about theora with text codecs...
Nov 11 15:30:39 <nessy> as a user, it's just a web app, right?
Nov 11 15:31:14 <denisc> Yep. But I worry about cross platform issues for users - so far only the oggplay stuff on Linux works for me.
Nov 11 15:31:20 <denisc> Everyone else here is mac
Nov 11 15:31:42 <nessy> ff3 should solve that, right?
Nov 11 15:32:10 <nessy> also, if you have somebody who would be keen to develop the IE plugin further, there will be code for them!
Nov 11 15:32:12 <denisc> I don't know - arkaiv falls apart badly accessed with ff3 so that's my next task
Nov 11 15:32:40 <nessy> is that because of mod_annodex?
Nov 11 15:32:43 <doublec> if you need help with ff3 stuff, let me know
Nov 11 15:32:55 <denisc> doublec:Cool - thanks!
Nov 11 15:33:01 <kfish> and if you need help with mod_annodex / oggz-chop stuff, let me know ;-)
Nov 11 15:33:17 <denisc> nessy: I guess mod_annodex is the culprit so will see what needs changing - thanks too kfish
Nov 11 15:33:38 <kfish> denisc, it may be worth updating libannodex svn, i rewrote the ogg chopping part of that recently
Nov 11 15:33:49 <denisc> OK - good one
Nov 11 15:34:04 <kfish> not fully finished yet, but should be better than it was before
Nov 11 15:34:20 <denisc> Hey - that's software though right? :)
Nov 11 15:34:25 <kfish> heh
Nov 11 15:34:36 * nessy makes a note to mention archaiv in list of projects next time
Nov 11 15:35:23 <denisc> nessy: thanks - but I was going to abandon it until it started getting interest from other staff, a PhD student and the eduaction faculty here...
Nov 11 15:35:34 <nessy> MikeS: what have you been up to these days?
Nov 11 15:36:03 <nessy> denisc: there's plenty of semi abandoned things on the list :)
Nov 11 15:36:38 <nessy> actually that reminds me... I should probably add your linguistic annotations to my list of use cases for text codecs...
Nov 11 15:36:42 <nessy> hmm....
Nov 11 15:37:35 <nessy> mostly the requirement is to be able to attach timing information on a per-word or per-syllable basis, right?
Nov 11 15:38:02 <denisc> Well not really - my needs are ethnomusicological so simple lyrics timed roughly works atm
Nov 11 15:38:16 <denisc> Linguists might need that though...
Nov 11 15:38:44 <nessy> ok, cool
Nov 11 15:38:50 <nessy> lyrics I have covered
Nov 11 15:39:36 <nessy> anything else before I close the meeting officially?
Nov 11 15:39:53 <nessy> (I mean: conversations can of course continue!!)
Nov 11 15:40:08 <kfish> nuthin' important from me!
Nov 11 15:40:16 <denisc> Me too
Nov 11 15:40:18 <nessy> everythin's important! ;)
Nov 11 15:40:33 <nessy> ok - meeting officially closed :)
