[1/5/2005 6:00 AM] So folks know, I can't be here past an hour:fifteen.... [1/5/2005 6:00 AM] if it makes you all feel better... most of us aussies were here at 8am due to a slight misundestanding :) [1/5/2005 6:00 AM] anyway, let's begin. [1/5/2005 6:00 AM] -->| Uraeus (~cschalle@213.170.47.91) has joined #xiphmeet [1/5/2005 6:00 AM] Anyone got a logging bot ready to go? [1/5/2005 6:00 AM] hey uraeus :) [1/5/2005 6:00 AM] hey ozone [1/5/2005 6:00 AM] I'm logging... [1/5/2005 6:01 AM] OK [1/5/2005 6:01 AM] There are people that don't log? [1/5/2005 6:01 AM] -->| wtay (~wtay@213.170.47.91) has joined #xiphmeet [1/5/2005 6:01 AM] I log, but not realtime. [1/5/2005 6:01 AM] OK, we've begun. Thanks for coming folks, I'm chairing the meeting today. [1/5/2005 6:01 AM] Starting with libogg2. [1/5/2005 6:02 AM] I've not yet gotten to reviewing/merging Arc's work. [1/5/2005 6:02 AM] However, we'd discussed it in the past, and I doubt there will be any reason to delay the merge beyond my own lack of time at the moment. [1/5/2005 6:03 AM] (and really, it's not 'lack of time' so much as 'many things to do, this is one'. It has been a productive month) [1/5/2005 6:03 AM] xiphmont: what does this mean for my other work with libogg2/OggStream [1/5/2005 6:03 AM] Once that merge happens, there will need to be a libogg1 compatability layer built on top, and a discussion of moving the other Xiph projects to using it. [1/5/2005 6:04 AM] Arc: in what sense? [1/5/2005 6:04 AM] should I continue holding for merger or just expect that it'll be merged without major API changes needed? [1/5/2005 6:04 AM] -->| j^ (~j@host-84-9-75-95.bulldogdsl.com) has joined #xiphmeet [1/5/2005 6:05 AM] If the API changes are what we discussed, they will be as is. [1/5/2005 6:05 AM] -->| karlH (~karl@82-38-34-147.cable.ubr02.brad.blueyonder.co.uk) has joined #xiphmeet [1/5/2005 6:05 AM] cool. [1/5/2005 6:05 AM] do you mean the namespace change? [1/5/2005 6:05 AM] <--| j^ has left #xiphmeet ("Client exiting") [1/5/2005 6:06 AM] OK, anyway, we were in agreement on all the points we discussed. [1/5/2005 6:06 AM] ok. [1/5/2005 6:06 AM] the namespace change is no big deal. [1/5/2005 6:06 AM] I don;t remember if the namespace changes were part of the current merge waiting to happen. [1/5/2005 6:06 AM] If they're not, they'll be done shortly after. [1/5/2005 6:06 AM] *nod* [1/5/2005 6:06 AM] you or me doing it? [1/5/2005 6:06 AM] i think i started but i dont remember where i started. [1/5/2005 6:06 AM] The only discussion to have is browbeating others to moving toward libogg2 native :-) [1/5/2005 6:07 AM] I'll do it. [1/5/2005 6:07 AM] cool. [1/5/2005 6:07 AM] (please give me a mapping file so i can appropriatly upgrade py-ogg2 - i need it for development) [1/5/2005 6:08 AM] Libogg2 is both zero-copy and lower memory usage (substantially lower). The usage is the same *except* who owns memory changes in several places. Most code will only change one or two calls; the bitpacker works the same way. [1/5/2005 6:08 AM] (This is meant for the other codec developers) [1/5/2005 6:08 AM] well and some of the processes are different, too. [1/5/2005 6:08 AM] Moving onto next project: I'm going to interject Postfish in here because it has seen some work and will be a major release in the next few months. [1/5/2005 6:09 AM] Postfish has a new reverb this week; much nicer than the old SWH reverb it replaces. It's based on Freeverb3 but adds predelay amd early-reflection control. [1/5/2005 6:09 AM] Thats' a small thing. [1/5/2005 6:10 AM] More importantly, I've begun the reorganization into postfish the app and libpostfish the programmatic library. [1/5/2005 6:11 AM] Among other things, the strict division will give multithreading support for free in the lib; the app will be able to scale to 32-way machines. [1/5/2005 6:11 AM] sorry for the ignorance, but postfish is similar to beast/bse ? [1/5/2005 6:11 AM] So, all you folks who sprung for a dual G5 can feel better about it :-) [1/5/2005 6:11 AM] It has one or two things in common with Beast if I recall correctly. [1/5/2005 6:12 AM] Beast is more about composition; Postfish is more about processing sound you already have. [1/5/2005 6:12 AM] ok thnx [1/5/2005 6:13 AM] Postfish is intended for complex inline filtering and/or mixdown of input audio. It does no synthesis, sequencing, etc. [1/5/2005 6:14 AM] OTOH, it has several unique restoration filters, and the implementation of the other filters are intended to be gold-standard grade. [1/5/2005 6:15 AM] the filters will be completely accessible from the lib ? [1/5/2005 6:15 AM] In any case, it will also offer them as a programmatic library, and I look forward to apps like Audacity using that to add to the 'standard' effect it offers as well as replacing other filters it already has (like EQ0 with a version that actually works right. And of course, having it available in a linear filtering package. [1/5/2005 6:16 AM] All filters will be in the lib. The application will be a GUI. [1/5/2005 6:16 AM] did you ever write up some rationale as to for example why not use ladspa ? [1/5/2005 6:16 AM] not ruffling feathers, just interested in knowing about your choice (outside the scope of the meeting) [1/5/2005 6:16 AM] (ditto for jack) [1/5/2005 6:17 AM] ladspa is a plugin architecture. You could easily make ladspa plugins with libpostfish. [1/5/2005 6:17 AM] And a beta tester is already jack-ifying Postfish. [1/5/2005 6:17 AM] awesome. [1/5/2005 6:18 AM] It's about the new filters and having a good UI that makes them easy to use. [1/5/2005 6:18 AM] how the filters are wrapped is secondary. [1/5/2005 6:18 AM] (it's not ladspa native as the ladspa api would be less efficient for massive multithreading) [1/5/2005 6:19 AM] (as well as missing other things like 'what's your latency?') [1/5/2005 6:19 AM] or at least, I've not seen any plugin use that API aspect if it does exist :-) [1/5/2005 6:19 AM] OK, moving on... [1/5/2005 6:20 AM] icecast. Do we have Icecast folks in attendance? [1/5/2005 6:20 AM] yeah [1/5/2005 6:20 AM] Oh, Karl is here. [1/5/2005 6:20 AM] go for it. [1/5/2005 6:20 AM] icecast 2.2 was release just before christmas [1/5/2005 6:21 AM] includes the theora and shoutcast DSP client support [1/5/2005 6:21 AM] so now we are doing video, NSV and theora streams have been tested [1/5/2005 6:22 AM] * illi would like to know of a test theora stream to iron out some bugs he has [1/5/2005 6:23 AM] I sure one can be set up, it's just a question on content [1/5/2005 6:23 AM] illi: ours are still up, as always [1/5/2005 6:23 AM] test pattern is fine :) [1/5/2005 6:23 AM] illi: if they're not, poke us [1/5/2005 6:23 AM] URI ? [1/5/2005 6:23 AM] I've got a link with some file streamed through icecast [1/5/2005 6:23 AM] http://mirror.fluendo.com [1/5/2005 6:23 AM] i never saw them there before... ok ! [1/5/2005 6:23 AM] the YP server has been updated to handle the video streams [1/5/2005 6:24 AM] i'd also like some icecast ones too, as there is bound to be some differences [1/5/2005 6:24 AM] illi: yeah, good idea. karlh ? [1/5/2005 6:24 AM] 217.195.20.1/video/shanecou.v196a56.ogg [1/5/2005 6:24 AM] How often is that up though ? [1/5/2005 6:24 AM] illi: Any eta for a new release of the filters? [1/5/2005 6:25 AM] easy enough to setup a stream, ezstream with the trunk of libshout can send static files [1/5/2005 6:25 AM] I'd guess end of feb is looking liekly... i'm kind of busy working, and i'd like to get theora streams and keyframe seeking in the next release [1/5/2005 6:25 AM] karlH : Run the install program and wha ? :) [1/5/2005 6:26 AM] illi: new clients on icecast usually start at the keyframe, so that should be no problem [1/5/2005 6:27 AM] true... the current bug is it geets confused when it sees high granule pos starting a stream... it tries to wait for that time and blocks [1/5/2005 6:27 AM] I can talk to you later about the ezstream side if you want [1/5/2005 6:27 AM] illi: I think you should ignore that - the first granule pos is the start of the stream, no ? [1/5/2005 6:27 AM] unless you seek to it [1/5/2005 6:28 AM] ie it's not distinguishing between it seeking to x (knowing it's x) and it just getting x... [1/5/2005 6:28 AM] it's not that hard to fix... i just need steams to test on [1/5/2005 6:29 AM] other than that, I see that oddsock is preparing another ezstream release, and I'm updating ices 2.0.1, which should be out today/tomorrow [1/5/2005 6:30 AM] Anything else? [1/5/2005 6:30 AM] the bug reports have been pretty thin on the ground, so we are taking that as good news [1/5/2005 6:30 AM] nothing more I think, back to you [1/5/2005 6:30 AM] OK then, Vorbis. [1/5/2005 6:31 AM] Not too much news there; I need to get the Neuros code into SVN and, while I'm at it, apply a few patches to mainline Tremor, replace mainline Tremor with the lowmem version... [1/5/2005 6:32 AM] xiphmont: is the lowmem version source-compatible with non-lowmem tremor? [1/5/2005 6:32 AM] Also I'm officially doing some oggenc optimization work for Mercora. [1/5/2005 6:32 AM] source compatable? as in same API? yes. [1/5/2005 6:32 AM] great [1/5/2005 6:32 AM] It has higher CPU complexity tho. [1/5/2005 6:32 AM] OTOH, it uses 1/10th the memory. [1/5/2005 6:33 AM] Some fun new thought problems have been panning out for Vorbis II too :-) I realized I'd missed an obvious possibility is structuring codebook cascades. [1/5/2005 6:34 AM] (in Vorbis I) [1/5/2005 6:34 AM] Also, looking into parametric codebooks. [1/5/2005 6:34 AM] what does that mean [1/5/2005 6:34 AM] 'codebooks that are both more effective and use less memory'. [1/5/2005 6:34 AM] xiphmont: hmm, how much more CPU? a significant amount? we're using tremor in a project at the moment, and we need all the CPU we can get [1/5/2005 6:35 AM] i'll take it with you offline though, since it's a bit off-topic [1/5/2005 6:35 AM] so VorbisI can be improved more than it's current implementation without breaking spec? [1/5/2005 6:35 AM] ozone: I'm not going to kill the normal version. CPU usage depends on the processor. I've not actually measured the diff on x86. [1/5/2005 6:35 AM] I should. [1/5/2005 6:35 AM] Arc: no, this is an algorthmic change. [1/5/2005 6:35 AM] ah ok [1/5/2005 6:35 AM] ozone: actually, you should measure and let me know ;-) [1/5/2005 6:36 AM] xiphmont: hehehe, sounds like a deal [1/5/2005 6:36 AM] Oh, on the announcement side of thing, Aoyumi released a new AoTuV version. [1/5/2005 6:36 AM] That's not Xiph.Org news, but folks might want to know. [1/5/2005 6:37 AM] I'm done with Vorbis news. Next: Theora. [1/5/2005 6:37 AM] nice of them to release it under a free license [1/5/2005 6:38 AM] thomasvs, others, anything new to announce? [1/5/2005 6:38 AM] yes [1/5/2005 6:38 AM] (or derf, illi, no rillian today) [1/5/2005 6:38 AM] illi: yes? [1/5/2005 6:39 AM] Ummm.. just mention ozone has ported a lot of my base libraries for max and unix... (C++) [1/5/2005 6:39 AM] also that we (mainly ozone) are working on server side seeking/recombination of ogg files [1/5/2005 6:39 AM] max = mac [1/5/2005 6:40 AM] Also... jsut my whats in store for this years... [1/5/2005 6:40 AM] that counts, yes :-) [1/5/2005 6:40 AM] xiphmont: well, source of our cortado applet has been released [1/5/2005 6:40 AM] xiphmont: it has vorbis and theora, but not yet speex [1/5/2005 6:40 AM] xiphmont: svn will be public too as soon as I get off my ass and move it [1/5/2005 6:40 AM] Just a small mention that we've released the code we've been working on for scalable network servers BTW [1/5/2005 6:41 AM] basically this years agenda... [1/5/2005 6:41 AM] Which is the project I was briefly alluding to before the meeting [1/5/2005 6:41 AM] keyframe seeking for theora... [1/5/2005 6:41 AM] xiphmont: I was wondering, wrt theora, what the current status is re: getting mmx optimizations working well [1/5/2005 6:41 AM] yeah, illi's windows stuff actually has a platform-neutral core [1/5/2005 6:41 AM] streaming of all codecs... [1/5/2005 6:41 AM] there's a lot of useful tools he has which are valuable outside of the directshow filters themselves [1/5/2005 6:41 AM] handling chains (or die trying) [1/5/2005 6:41 AM] (i'll wait a bit for illi and ozone first) [1/5/2005 6:42 AM] e.g. an Ogg validation tool, a chain splitter [1/5/2005 6:42 AM] (chain splitter == useful for listening to streaming vorbis radio; it'll split chains and rename them appropriately based on the filename) [1/5/2005 6:42 AM] dump tools, chain info yada yada [1/5/2005 6:42 AM] based on the metadata in the radio stream, i mean [1/5/2005 6:42 AM] yeah, those kind of things. [1/5/2005 6:42 AM] thomasvs: mmx is not something I;d worked on. Beyond the patch submitted to the list, I'm not aware of much/any MMX work going on... but rillian would know better. [1/5/2005 6:43 AM] xiphmont: ok, will poke him again [1/5/2005 6:43 AM] also support vfw (video for windows) for theora so it can be used in the usual editing suites... [1/5/2005 6:43 AM] probably the first significant addition to his set of tools is mod_oggchef, which is an apache 2 module that enables server-side seeking of ogg files via timed URLs [1/5/2005 6:43 AM] Project is here: http://sourceforge.net/projects/kamaelia [1/5/2005 6:43 AM] We'll be releasing the code we use to wrap vorbis up into something convenient to wrap as a submodule [1/5/2005 6:43 AM] but anyway, if you're looking for alternative libraries to, say, libogg, and aren't afraid of C++, have a look at the oggdsf/unix directory in the xiph SVN trunk [1/5/2005 6:43 AM] and my big goal is to have installed base of 1mil + (hoping for 10... but we'll see by years end) [1/5/2005 6:44 AM] illi: how do you measure achievement of that goal ? [1/5/2005 6:44 AM] i'll post to ogg-dev about this soon. that's all from me ... [1/5/2005 6:45 AM] that was more than I expected to hear about with rillian absent, and it will take me years to demultiplex it all :-) [1/5/2005 6:45 AM] well obviously as there's no spyware... i can't be sure... but i do know how many people download them [1/5/2005 6:45 AM] heh :) [1/5/2005 6:45 AM] illi: obviously the solution is to add spyware, to enable an accurate count [1/5/2005 6:45 AM] furthermore, on the GStreamer level, wim is hacking on his redesigned branch, focusing on ogg stuff first [1/5/2005 6:45 AM] (BTW the purpose of having that code written in the first place was to allow us to test the streaming protocols test server we're writing) [1/5/2005 6:46 AM] hovering around 600-1500 per day (not including the spike at release times) [1/5/2005 6:46 AM] theora video scrubbing is very zippy, chained oggs work fine. I hope to keep ogg stuff as first class citizens [1/5/2005 6:46 AM] thomasvs: ooh, nice. [1/5/2005 6:46 AM] <--| mhrd has left #xiphmeet [1/5/2005 6:47 AM] Hm, no Jean-Marc. Anyone else have Speex news? [1/5/2005 6:47 AM] I have this: Work on the Speex parrot continues :-) [1/5/2005 6:47 AM] And there is also a Flac Frog. [1/5/2005 6:47 AM] A vorbis vulture ? [1/5/2005 6:48 AM] * illi has no idea what parrot frog is ! [1/5/2005 6:48 AM] Anyway, I'll throw the floor open to speex news and flac news if folks have some, but Josh isn't here either, and I frankly don't blame him given the hour. [1/5/2005 6:48 AM] Oh, we have a new fish too. I'll get to show that around soon. [1/5/2005 6:48 AM] cool [1/5/2005 6:48 AM] do we have codec icons yet? [1/5/2005 6:48 AM] (something for Writ, prehaps? hehe) [1/5/2005 6:49 AM] Speex/flac. Vorbis will still use the fish. [1/5/2005 6:49 AM] No progress on Theora/Writ yet. [1/5/2005 6:49 AM] ok. [1/5/2005 6:49 AM] I have an update on Writ. [1/5/2005 6:49 AM] oh icons... it's clear now :) [1/5/2005 6:49 AM] Melissa is still toying with the cat/TV idea for Theora, but hasn;t found one that works yet. [1/5/2005 6:49 AM] No racoon for Theora? [1/5/2005 6:49 AM] it was a cat. [1/5/2005 6:49 AM] Theora has always been a cat [1/5/2005 6:49 AM] o_O [1/5/2005 6:49 AM] I suggested a cat trapped *in* a tv, but... [1/5/2005 6:49 AM] the striped tail and weird eye made it look like a racoon [1/5/2005 6:50 AM] what about a cat eating a TV? [1/5/2005 6:50 AM] FWIW, we still use "the fish" for any xiph-related codec as an overlay [1/5/2005 6:50 AM] that'd be appropriate for australia. australian cats eat TVs all the time [1/5/2005 6:50 AM] see, these things all do not scale down. [1/5/2005 6:50 AM] would be nice to have some codec art finished sometime soon [1/5/2005 6:50 AM] xiphmont: dont get animal welfare on us, for gods sakes [1/5/2005 6:50 AM] thomasvs: the fish is finished. [1/5/2005 6:50 AM] the last thing we need is the SPCA or PETA going after us [1/5/2005 6:50 AM] and the fish will remain the overarching icon for all things xiph. [1/5/2005 6:51 AM] where's your sense of adventure, arc [1/5/2005 6:51 AM] cat's a delicacy down here... [1/5/2005 6:51 AM] It's not like he's going to pick a cat being chopped in half for Theora, is he? [1/5/2005 6:51 AM] ozone: I used to be a card carrying member of PETA. I know what they're capable of. trust me, not what we want to deal with. [1/5/2005 6:51 AM] ok, back on topic. [1/5/2005 6:51 AM] Arc: writ [1/5/2005 6:51 AM] Writ? [1/5/2005 6:52 AM] Writ is done. [1/5/2005 6:52 AM] People for the Eating of Tasty Animals? [1/5/2005 6:52 AM] um, the problem is implementation [1/5/2005 6:52 AM] spec/code? [1/5/2005 6:52 AM] ah. [1/5/2005 6:52 AM] In the wiki, isn't it? [1/5/2005 6:52 AM] see, that's the thing. i cant figure out how a library would be helpful [1/5/2005 6:52 AM] the only helpful thing it could do is repeat long phrases [1/5/2005 6:52 AM] http://wiki.xiph.org/OggWrit [1/5/2005 6:52 AM] Atamido: yes [1/5/2005 6:52 AM] well, without code, the spec doesn't do much.... [1/5/2005 6:52 AM] there's example code in py-ogg2 [1/5/2005 6:53 AM] xiphmont: exactly. so I need help getting some media player to actually use it, or some encoder/muxer to work with it [1/5/2005 6:53 AM] there, I am stuck. [1/5/2005 6:53 AM] Maybe a library would offer multiple interfaces, one for each language? [1/5/2005 6:53 AM] Atamido: offline. [1/5/2005 6:54 AM] Arc: research how other subtitle systems integrate with a player? [1/5/2005 6:54 AM] so I'm sending out an advertisement. if anyone knows anyone who can work on the theora plugins for any media player, I need their help. [1/5/2005 6:54 AM] I assume it's compartmentalized code elsewhere. [1/5/2005 6:54 AM] I can help with this, but I dont have the time myself to learn media player stuff right now [1/5/2005 6:54 AM] Arc: gst is waiting for you [1/5/2005 6:55 AM] thomasvs: we'll talk after meet. thanks :-) [1/5/2005 6:55 AM] thats it. [1/5/2005 6:55 AM] Spiffy. [1/5/2005 6:56 AM] I have only one last thing to discuss. [1/5/2005 6:56 AM] (didn;t add to agenda, sorry) [1/5/2005 6:56 AM] I have been considering dropping the codec projects back to officially making source-only releases. [1/5/2005 6:57 AM] This has obvious benefits and liabilities. [1/5/2005 6:57 AM] the biggest benefit is that we won't constantly be stalled on building/testing binaries. [1/5/2005 6:58 AM] The liability is that binary downloads from the main xiph site would either lag source releases or mostly disappear. [1/5/2005 6:58 AM] hm [1/5/2005 6:58 AM] are there any measurements on how popular the binary downloads are? [1/5/2005 6:58 AM] However, the codecs/libs today are mostly distributed bu distros and other apps. [1/5/2005 6:59 AM] Someone could do log analysis of the web server logs to see how big it is. [1/5/2005 6:59 AM] I'm talking about the codecs cores, not things like Ogg DSF [1/5/2005 6:59 AM] Unless illi had good reason it would make sense for that too. [1/5/2005 6:59 AM] Are you including things like the oggenc/flac executables? [1/5/2005 7:00 AM] Possibly; most people either get those from distros or build from source. [1/5/2005 7:00 AM] Unless illi had good reason it would make sense for that too. ????? [1/5/2005 7:00 AM] But I don;t know exactly how vast that majority is. [1/5/2005 7:00 AM] True for Unix/Linux. Not true for Windows, I think. [1/5/2005 7:01 AM] illi: forgive my grammar. [1/5/2005 7:01 AM] malcolm: I agree. [1/5/2005 7:01 AM] I disagree & I know I am not alone Vorbis 1.1 was a mess & people use Hydrogen shitty _impulse_trigger build without even knowing it's not official ... [1/5/2005 7:01 AM] However, Win users actually tend to use third party tools. [1/5/2005 7:01 AM] xiphmont: I couldn't parse it at all... have i like missed some lines or somethign ? [1/5/2005 7:01 AM] Speaking as a user of both, I don't need Linux binaries, but not providing Windows binaries would really be annoying. [1/5/2005 7:02 AM] xiphmont: yes and no. they currently use third-party tools because there's not much in the way of directshow support for ogg [1/5/2005 7:02 AM] illi: "I'm not suggesting the same for the direct show filters, unless for some reason, illi thinks it would be a good idea to only release it as source'. I don;t think you think that :-) [1/5/2005 7:02 AM] i'd wager that illi's oggdsf stuff will be far more dominant soon, if it's not already [1/5/2005 7:02 AM] Building stuff on Windows is neither common, nor terribly easy. [1/5/2005 7:02 AM] Well... about 5000:1 binary to source [1/5/2005 7:03 AM] (Vorbis 1.1: is there even a windows build?) [1/5/2005 7:03 AM] well, part of it is none of the core folks are Win maintainers. 1.1 is delayed because of Windows. It's holding up all the other platforms too. [1/5/2005 7:03 AM] Not a lot of people have the DXSDK installed. [1/5/2005 7:03 AM] So we can stall indefinitely, or keep doing what we do well; improving the source. [1/5/2005 7:03 AM] (and at 300 megs... phwoar!) [1/5/2005 7:03 AM] xiphmont: i'm prepared to be the windows maintainer for vorbis/theora [1/5/2005 7:03 AM] (and other codecs, if necessary) [1/5/2005 7:04 AM] atamido : Building stuff on windows is vastly easier... Right-click/build all [1/5/2005 7:04 AM] ozone: much of the difficulty is that the win32 people we've had in the past were somewhat... high maintainence. Either easily distracted, or unwilling to move forward without supervision. [1/5/2005 7:04 AM] atamido : But only if you distribute it in a form ammenable to that [1/5/2005 7:04 AM] if linux binaries are still considered, I'd be more than happy to be doing redhat/fedora packages [1/5/2005 7:04 AM] xiphmont: i'm up for it [1/5/2005 7:04 AM] I don;t want a volunteer so much as a person raring to take that aspect of the project over and do it 'good 'n' hard'. [1/5/2005 7:04 AM] pem: I've made a FC3 package for ICecast still not avalaible on the website [1/5/2005 7:05 AM] xiphmont: i'm nubile and raring. what else is there to say? :) [1/5/2005 7:05 AM] I've also suggested to some Fedora mainteners to include icecast on the main repos [1/5/2005 7:05 AM] * thomasvs cheers on ozone [1/5/2005 7:06 AM] OK, so... it's clear there's discussion to be had. My desire is that we don't let the quagmire of release enineering sap what we actually used to do very well: hack on cool projects. Releases have been sucking the life out of us. That has to stop. This was one proposed solution, so let's continue the discussion on the lists. [1/5/2005 7:06 AM] * HackRip wants Case as Win32 guy ... if ever you read the log Case ;) [1/5/2005 7:06 AM] ok [1/5/2005 7:06 AM] ozone: I will take you up on your offer! [1/5/2005 7:06 AM] illi: You are assuming that people have something installed that _can_ build sources. [1/5/2005 7:06 AM] Whoever grabs this and moves forward wins the football. [1/5/2005 7:07 AM] xiphmont: sure. i'll post to the vorbis & theora (dev) lists about it, then [1/5/2005 7:07 AM] ozone is the build masta ! [1/5/2005 7:07 AM] * illi doesn't have the patience to maintain in two places [1/5/2005 7:07 AM] ozone: There's more to say, we'll handle it later on. I need to get a status dump from Rillian. [1/5/2005 7:07 AM] xiphmont: ok, feel free to email me then [1/5/2005 7:07 AM] illi: 3 places, you forget about mac os x :) [1/5/2005 7:08 AM] ozone: as a start, do this tho: grab sources and build. If it doesn;t build and test functional, begin determining why :-) [1/5/2005 7:08 AM] ozone, macosx is kinda like bothering with supporting amiga with binaries, so he did not forget :) [1/5/2005 7:08 AM] i think we at least pass that test right now, but II want someone to know *for sure* :-) [1/5/2005 7:08 AM] no i meant my own configuration (ie fully integrated) and seperately... not platforms [1/5/2005 7:08 AM] xiphmont: will do tonight [1/5/2005 7:09 AM] thanks. [1/5/2005 7:09 AM] Uraeus :P [1/5/2005 7:10 AM] That point is now handled within the scope of the meeting. [1/5/2005 7:10 AM] Any other items not on the agenda before we adjourn? [1/5/2005 7:11 AM] hep ! [1/5/2005 7:11 AM] RTP status report [1/5/2005 7:11 AM] hum, RTP first [1/5/2005 7:11 AM] yes, please. [1/5/2005 7:11 AM] Ok, An -04 update of the Vorbis draft has been sent to the IETF [1/5/2005 7:11 AM] ozone : you might want to start here... www.illiminable.com/th/XiphWinSDK.zip [1/5/2005 7:11 AM] save you mucho time [1/5/2005 7:12 AM] i tried maintaining two seperate tree for a while... but it shat me [1/5/2005 7:12 AM] illi: cheers [1/5/2005 7:12 AM] Changes covered moving the config packets from RTCP to in-band as well as via SDP [1/5/2005 7:12 AM] (But it all builds, and the enc/dec tools) [1/5/2005 7:12 AM] Work is progressing on a reference Vorbis RTP server [1/5/2005 7:12 AM] ozone: start with CVS, when that doesn;t work, look at illi's help ;-) [1/5/2005 7:12 AM] he'll just need to copy the vcproj files... [1/5/2005 7:13 AM] philk: is working code a required part of the draft or just a good idea? [1/5/2005 7:13 AM] xiphmont: that's the plan [1/5/2005 7:13 AM] I sent an update of the Speex I-D to the chaps a week or so ago, no word back [1/5/2005 7:13 AM] the chaps being.... [1/5/2005 7:13 AM] xiphmont: The code needs to be done at some point of the standardisation process, the earlier the better [1/5/2005 7:14 AM] The BBC have said they will make implementations :) [1/5/2005 7:14 AM] OK, so it is a requirement [1/5/2005 7:14 AM] :-) [1/5/2005 7:14 AM] Next on the list is the Theora draft [1/5/2005 7:14 AM] An initial version should be done by the end of the month [1/5/2005 7:15 AM] Is the theora spec document providing the start? I'm sure you've been over this with rillian... [1/5/2005 7:15 AM] Yup, i need to run through some of the finer points [1/5/2005 7:15 AM] We're now at the point I must go... [1/5/2005 7:15 AM] Done [1/5/2005 7:16 AM] was 'hem' an agenda item or a throat clearing noise? [1/5/2005 7:16 AM] :) [1/5/2005 7:16 AM] I just wanted to talk about the docs, mostly the icecast docs [1/5/2005 7:16 AM] er, 'hep' [1/5/2005 7:16 AM] setting up a translation/help team or effort [1/5/2005 7:16 AM] *Atamido wanted to poke xiphmont about website stuff, but will wait until later? [1/5/2005 7:16 AM] Ah yes! [1/5/2005 7:17 AM] Atamido: yes. I forwarded mail to you before the meeting [1/5/2005 7:17 AM] If you do not have more time, we could talk about that via mail [1/5/2005 7:17 AM] pem: go ahead. I need to leave the meeting tho. [1/5/2005 7:17 AM] xiphmont: What email address? [1/5/2005 7:17 AM] no,no, it was on the agenda. [1/5/2005 7:17 AM] Atamido: apparently the wrong one. Will resend in a bit. [1/5/2005 7:18 AM] Anyway, folkss hould adjourn when the doc discussion is done. I really must run. [1/5/2005 7:18 AM] Are there any plans for doc-translations for projects other than icecast and flac-russian? [1/5/2005 7:18 AM] thanks for coming, I look forward to seeing the final log ;-) [1/5/2005 7:18 AM] |<-- xiphmont has left irc.freenode.net ("He travels fastest who travels alone") [1/5/2005 7:19 AM] Doh... [1/5/2005 7:19 AM] Atamido: it was one of my ideas [1/5/2005 7:19 AM] What is the global plan of your idea? [1/5/2005 7:19 AM] to launch an official translation team [1/5/2005 7:20 AM] and call for help [1/5/2005 7:20 AM] You can see it on the list [1/5/2005 7:20 AM] but he travels "alone" who travels fastest ... [1/5/2005 7:20 AM] Many users are willing to help [1/5/2005 7:20 AM] but are (so am I) unable to code [1/5/2005 7:20 AM] I'm a french guy [1/5/2005 7:20 AM] I'm sure that some users are german. [1/5/2005 7:20 AM] etc. [1/5/2005 7:20 AM] Then please don't make any docs in english. [1/5/2005 7:21 AM] A doc in english is good enough for coders [1/5/2005 7:21 AM] Is this a plan for translating all docs from all projects? [1/5/2005 7:21 AM] Yep [1/5/2005 7:21 AM] No that fast, but to begin [1/5/2005 7:21 AM] Any idea how to do the infrastructure? [1/5/2005 7:21 AM] I've been working as a translator for Kde [1/5/2005 7:21 AM] |<-- HackRip has left irc.freenode.net () [1/5/2005 7:22 AM] Where are the docs put? [1/5/2005 7:22 AM] Working on KMail [1/5/2005 7:22 AM] How do you access a particular doc in another language? [1/5/2005 7:22 AM] At first we can simply begin on the Xiph wiki [1/5/2005 7:22 AM] list the docs [1/5/2005 7:22 AM] make a first roadmap [1/5/2005 7:23 AM] even if a soft is realesed in another version, most of the doc is still valuable [1/5/2005 7:23 AM] for me the infrastructure is not that important [1/5/2005 7:23 AM] An example of one way of doing the multilingual, look at the picture flags at the top of this page. v [1/5/2005 7:23 AM] http://www.matroska.org/contact/index.html [1/5/2005 7:24 AM] Only translated pages are highlighted. [1/5/2005 7:24 AM] The most important point is that Xiph, or the icecast website, officialy calls for help [1/5/2005 7:24 AM] Yep. [1/5/2005 7:24 AM] Good idea [1/5/2005 7:24 AM] Well, if people can't easily access the translated pages, it is moot. [1/5/2005 7:25 AM] Yes, of course [1/5/2005 7:25 AM] But we first need to begin translating them [1/5/2005 7:25 AM] And we need to ask for some help [1/5/2005 7:25 AM] Spend a day figuring out where to put the and how to make them accessible before asking for help translating. [1/5/2005 7:25 AM] Otherwise you end up with some real messes. [1/5/2005 7:26 AM] :/ [1/5/2005 7:26 AM] Sorry but a few seconds ago the meeting was up to finish without tallink about that point [1/5/2005 7:26 AM] I've already send a mail on the icecast mailing list without any response [1/5/2005 7:27 AM] I can make you a mail with the project [1/5/2005 7:27 AM] and how to begin, how to do it [1/5/2005 7:27 AM] I'm not sure at all that thinkink at "how will look the translated website" is the good way to start [1/5/2005 7:28 AM] Can you give me your email address? [1/5/2005 7:28 AM] Well, the website is relative close to being redone, so I don't think you could do a lot of that. [1/5/2005 7:29 AM] It is a question though if the plan is to have all of the webpages slowly translated to other languages, like the site linked above. [1/5/2005 7:29 AM] Or just the docs. [1/5/2005 7:29 AM] And where are the docs going to be put. [1/5/2005 7:30 AM] Are there plans to merge the docs to one central location? [1/5/2005 7:30 AM] If you plan on starting translations in the wiki, what page naming scheme do you want to use so people can find the translations? [1/5/2005 7:30 AM] Sorry but I must be in a bad mood but I've got the feeling you're not really interested [1/5/2005 7:30 AM] Little things like that. [1/5/2005 7:31 AM] <--| ozone has left #xiphmeet [1/5/2005 7:31 AM] No, I just don't want to have to clean up a bunch of pages after they are written. [1/5/2005 7:31 AM] I may look silly and stupid but I was not going to start translatid webpages naked on a field and them send them via email. [1/5/2005 7:31 AM] <--| karlH has left #xiphmeet ("Client Exiting") [1/5/2005 7:32 AM] How and where you translate is all up to you, I'm just concerned about how/where they are stored. [1/5/2005 7:32 AM] okay [1/5/2005 7:32 AM] I don't tend to wear pants inside my apartment, but I'm pretty firm on them in the field. [1/5/2005 7:33 AM] Anyway, either webmaster at xiph.org or atamido at gmail.com [1/5/2005 7:33 AM] Ok enough said, have a nice day. [1/5/2005 7:33 AM] <--| pem has left #xiphmeet