Logs for jdev

Show join/part/nick changes:

[05:39:54] * bear left the chat.
[06:31:59] * ermine joined the chat.
[07:13:33] * Alex joined the chat.
[07:22:45] * Guus joined the chat.
[07:24:47] * Tobias joined the chat.
[07:32:33] * tkoski joined the chat.
[07:38:09] * MattJ joined the chat.
[07:45:03] * Guus left the chat.
[07:45:05] * Guus joined the chat.
[07:52:51] * teo left the chat.
[07:52:52] * teo joined the chat.
[08:17:25] * jprieur joined the chat.
[08:21:24] * luca tagliaferri joined the chat.
[08:25:02] * teo left the chat.
[08:25:02] * teo joined the chat.
[08:39:51] * jprieur left the chat.
[08:56:53] * petermount joined the chat.
[09:00:20] * scippio joined the chat.
[09:03:29] * Treebilou joined the chat.
[09:04:34] * Treebilou left the chat.
[09:04:36] * Treebilou joined the chat.
[09:23:08] * luca tagliaferri_ joined the chat.
[09:25:55] * luca tagliaferri_ left the chat.
[09:28:11] * luca tagliaferri_ joined the chat.
[09:28:44] * luca tagliaferri_ left the chat.
[09:34:24] * Alex left the chat.
[09:34:25] * Alex joined the chat.
[09:48:50] * Alex left the chat.
[09:48:50] * Alex joined the chat.
[09:49:54] * Alex left the chat.
[09:50:03] * Alex joined the chat.
[10:13:26] * deryni left the chat.
[10:29:33] * mlundblad joined the chat.
[10:31:17] * dwd joined the chat.
[10:32:11] <dwd> !rfc 5802
[10:32:43] <dwd> I was just curious if a bot would respond. Now I see there isn't one about.
[10:33:52] <dwd> RFC 5802 Title: Salted Challenge Response Authentication Mechanism (SCRAM) SASL and GSS-API Mechanisms
[10:33:57] <dwd> Finally out.
[10:34:55] * Florob joined the chat.
[10:38:14] <Tobias> dwd: yeah..finally :)
[10:47:35] * Guus left the chat.
[11:04:17] * scippio left the chat.
[11:47:48] * Guus joined the chat.
[12:09:37] * Alex left the chat.
[12:09:38] * Alex joined the chat.
[12:26:40] * alkino joined the chat.
[12:31:43] * Alex left the chat.
[12:31:43] * Alex joined the chat.
[12:40:35] * mlundblad_netbook joined the chat.
[12:41:36] * mlundblad_netbook left the chat.
[12:45:03] * mlundblad_netbook joined the chat.
[12:47:06] * mlundblad_netbook left the chat.
[12:51:12] * mlundblad_netbook joined the chat.
[13:02:06] * test joined the chat.
[13:02:29] * test left the chat.
[13:04:08] * alkino left the chat.
[13:10:57] <dwd> Just submitted: http://dave.cridland.net/xeps/mobile.html
[13:11:14] * deryni joined the chat.
[13:12:36] * mlundblad_netbook left the chat.
[13:18:14] * mlundblad_netbook joined the chat.
[13:20:01] <Tobias> dwd: the usual client session receives way more than it sends, right?
[13:20:25] <MattJ> Yes, mostly through presence
[13:23:32] <dwd> Tobias, Unless you're being really talkative.
[13:24:11] <Tobias> dwd: you know average times after what idle time devices lower the radio state?
[13:24:33] <deryni> Or sending geoloc info from a plane?
[13:25:48] <Tobias> deryni: when that plane has other flying planes on its roster it might receive more geoloc than it sends :)
[13:26:05] <deryni> Air-traffic control via pep?
[13:26:10] <dwd> Tobias, It's not the device's choice - the timeouts are set by the network operator.
[13:26:27] <dwd> deryni, Air traffic control happens by X.400. Are you scared yet?
[13:26:52] <deryni> I imagine I might be if I knew what that was. =)
[13:27:42] <dwd> Tobias, So, anyway, I just measured my phone, basically, but I know the DCH timeout is usually a few seconds, and the FACH timeout can be anything from a few seconds to a couple of minutes.
[13:28:08] <deryni> One of my coworkers at my last job used to follow up silly/stupid stories by reminding everyong that he used to be a programmer for NASA.
[13:28:27] <Tobias> /me looks up DCH and FACH
[13:28:29] <Tobias> :)
[13:28:34] <deryni> And then commenting on how scary he still finds that.
[13:29:56] <Tobias> dwd: how dare you use abbreviations that aren't notable for wikipedia?
[13:29:58] <Tobias> :)
[13:30:03] * alkino joined the chat.
[13:30:26] <dwd> Tobias, What do you mean?
[13:30:35] <Tobias> FACH and DCH?
[13:30:45] <dwd> Tobias, All mobile phones communicate for low packet sizes using the Chilean Air Force.
[13:31:10] <Tobias> okay.. :)
[13:31:11] <dwd> Tobias, Larger bandwidth is provided by an obscure Hong Kong conglomerate.
[13:33:36] * mlundblad_netbook left the chat.
[13:33:39] * mlundblad_netbook joined the chat.
[13:34:09] * Lele joined the chat.
[13:34:28] * Lele left the chat.
[13:34:50] <Tobias> dwd: i don't remember right now but can memory level and compression level be adjusted in the middle of a stream?
[13:35:29] * tofu joined the chat.
[13:35:40] <Tobias> so that servers could adjust them for clients whose disco response states that they are a mobile client
[13:45:27] * Zash joined the chat.
[13:47:07] <dwd> Tobias, What, boost them to maximum for mobile clients?
[13:47:44] <Tobias> yeah..for example, to transfer as little as possible
[13:47:53] <dwd> It's an interesting idea, but I don't think you can do that.
[13:47:56] * alkino left the chat.
[13:48:02] <dwd> Have a look in zlib.h, though.
[13:48:18] <Kev> dwd: you could during stream setup, though.
[13:48:22] <Kev> No?
[13:48:38] <Tobias> Kev: but do you know already if it's a mobile client at that point?
[13:48:54] <Kev> Tobias: you could do, if you had a stream feature to say that you were.
[13:49:08] <dwd> The only thing that sets the alarm bells ringing is that if there an "I'm a mobile client, compression to warp factor nine, Sulu" switch, everyone will press it.
[13:49:25] <Kev> Right.
[13:49:45] <dwd> Or worse - consider it a "please chew up 300k of your valuable memory" button.
[13:49:47] <Tobias> Kev: i was thinking about those http://xmpp.org/registrar/disco-categories.html#client
[13:49:52] <dwd> Then consider who'll push it.
[13:50:00] <Kev> And then they won't get their presence until they send messages, and might think better of it:)
[13:50:06] <Kev> Tobias: I know.
[13:50:37] <dwd> Kev, That might be sufficiently evil. The DOSsers will still use it, though.
[13:51:04] <Tobias> the DOSers can also DOS you by sending malicious stanzas, etc.
[13:51:37] <dwd> But as I say, I found that dropping the memory cost quite considerably - say, wbits of 13, low memLevel - would drop compression to 25% instead of 20%, and it's not the critical side of the compression codec, then.
[13:52:18] <Tobias> dwd: would be still reasonable for a server to spend more memory/time on compression for mobile clients compared to desktop clients or other servers
[13:52:36] * lastsky left the chat.
[13:53:15] <dwd> Tobias, Yup. Still pales next to the problem that TLS stacks don't expose the compression knobs to the application.
[13:54:36] * mlundblad_netbook left the chat.
[13:56:08] <Tobias> dwd: yeah..so the servers wanting to do some clever thing would need to avoid compression during TLS. that should be possible in all TLS APIs
[13:56:44] <dwd> Tobias, That said, on the downlink, the primary implications of sending a bit more data are possibly finanical cost, and you may cause DCH to take longer to time out. There's no direct power saving from saving an octet or two.
[13:57:42] <dwd> Tobias, Plus, I couldn't extract any useful figures from experimentation with XMPP data on how the timeouts are affected by data volume because they were consistent with me, so I suspect you really have to crank up the data before you see it move off the minimum timeout.
[13:58:07] <Tobias> aha..okay
[13:58:19] <Tobias> i haven't made any practical tests myself :)
[13:59:34] * Alex left the chat.
[14:00:50] <Tobias> dwd: is it true that if you buffer more stanzas up before sending at server side you have a better compress ratio?
[14:05:36] <dwd> Tobias, Nope.
[14:05:56] <dwd> Tobias, I thought it might help, but aside from reducing TCP/TLS overhead, it doesn't.
[14:06:24] <Tobias> okay...i thought the larger the data set the more common bits it might find
[14:06:55] * jprieur joined the chat.
[14:06:57] * jprieur left the chat.
[14:07:53] <dwd> Tobias, No, it effectively has the larger data set at all times.
[14:08:19] <dwd> Tobias, As in, there's an internal buffer it uses for that. Which is why it's quite chunky on memory.
[14:08:27] <Tobias> ahh..okay
[14:09:35] <Tobias> dwd: are you going to post your proto-xep to jdev, standards and mobile lists?
[14:12:01] <dwd> Could do. I've sent it Peterwards, though, so I'd imagine he will.
[14:17:34] * Florob left the chat.
[14:18:05] <Tobias> dwd: ah..okay
[14:41:22] * alkino joined the chat.
[15:14:15] * teo left the chat.
[15:16:45] * alkino left the chat.
[15:18:54] * stpeter joined the chat.
[15:26:40] * john left the chat.
[15:29:50] * scippio joined the chat.
[15:32:10] * Guus left the chat.
[15:38:00] <stpeter> dwd: yes, received, I'll process that now
[15:39:04] * Tobias left the chat.
[15:39:05] * Tobias joined the chat.
[15:39:13] <dwd> stpeter, Thanks. After the IETF deadline, too. ;-)
[15:39:50] <stpeter> heh
[15:39:51] <stpeter> yeah
[15:39:56] <stpeter> no deadlines at the XSF!
[15:45:17] * hawke joined the chat.
[15:45:55] <stpeter> done
[15:48:36] * Tobias left the chat.
[15:48:55] * Tobias joined the chat.
[15:50:35] <dwd> stpeter, Thanks.
[16:02:23] <dwd> http://www.isode.com/whitepapers/xmpp-performance-constrained.html may be of interest. I normally don't pimp our whitepapers, but this one I found quite surprising.
[16:03:46] * teo joined the chat.
[16:05:29] <Tobias> dwd: compared to IRC, is IRC that much used in constrained environments?
[16:06:24] <Tobias> i thought IRC was only used by old free software people :)
[16:06:35] <dwd> And military over SatCom.
[16:08:33] * alkino joined the chat.
[16:14:46] <dwd> But yeah, IRC usage in military ruins my gag about tactical chat degenerating into "I totally pwned that guy!!! LOL!!!1!"
[16:17:48] <Tobias> heh :)
[16:26:07] * tkoski left the chat.
[16:27:07] <stpeter> /me fights his way through an overfull inbox
[16:42:08] * jkhii joined the chat.
[16:56:20] * Will joined the chat.
[16:57:36] * mcepl left the chat.
[16:57:42] * petermount left the chat.
[16:57:52] <Will> dwd: very interesting read!
[16:58:30] * mcepl joined the chat.
[17:00:15] * will.thompson joined the chat.
[17:01:22] * will.thompson left the chat.
[17:16:21] * Florob joined the chat.
[17:18:38] <stpeter> hmm
[17:19:22] <stpeter> "Optimized XMPP"??? Have British spelling conventions lost favour with the Isode team?
[17:20:03] <Kev> For some reason I've not yet worked out, we use Oxford English.
[17:20:11] <stpeter> also http://www.isode.com/xmpp-constrained-bandwidth.html yields a 404
[17:20:47] <stpeter> but that would, certainly, yield "optimised", no?
[17:21:06] <stpeter> the zed deserves to be included only in words derived from Greek, not Latin
[17:21:52] <stpeter> http://www.isode.com/whitepapers/xmpp-constrained-bandwidth.html is the correct URL, naturally
[17:22:51] <stpeter> and it would be good to document more efficient stream setup as outlined on the standards@ list some time ago (plus use of TLS reconnect or whatever it's called)
[17:23:25] <Kev> Right -although what this is doing isn't 'more efficient' stream setup.
[17:23:28] <Kev> It's no stream setup.
[17:23:44] <stpeter> yep, I didn't say they were the same
[17:23:48] <Kev> Or I guess 'out of band stream setup, to be accurate.
[17:24:24] <stpeter> only that we need to document improvements to stream setup (using standardized methods) so that we can make the normal setup process as optimised as possible
[17:26:41] <dwd> Kev, we ought to actually discuss that on the list, I suppose.
[17:27:29] * waqas joined the chat.
[17:40:30] * luca tagliaferri left the chat.
[17:43:12] * jugg left the chat.
[17:48:07] * Will left the chat.
[18:01:10] * Xificurk joined the chat.
[18:36:14] * teo left the chat.
[18:36:15] * teo joined the chat.
[18:58:44] * bear joined the chat.
[18:58:55] * smoku joined the chat.
[19:19:15] * jprieur joined the chat.
[19:19:30] * jprieur left the chat.
[19:26:41] * mcepl left the chat.
[19:27:22] * mcepl joined the chat.
[19:33:43] * Link Mauve joined the chat.
[19:54:40] * mcepl left the chat.
[20:19:22] * evilotto joined the chat.
[20:26:11] * Treebilou left the chat.
[20:29:21] * petermount joined the chat.
[20:41:58] * luca tagliaferri joined the chat.
[20:55:18] * hawke left the chat.
[20:55:18] * hawke joined the chat.
[21:08:57] * petermount left the chat.
[21:19:33] * Kev left the chat.
[21:19:41] * Kev joined the chat.
[21:20:37] * Tobias left the chat.
[21:38:17] * ermine left the chat.
[21:39:57] * johnny left the chat.
[21:40:46] * johnny joined the chat.
[21:43:31] * mlundblad_netbook joined the chat.
[21:50:29] * johnny left the chat.
[21:51:17] * johnny joined the chat.
[22:03:11] * Florob left the chat.
[22:05:32] * johnny left the chat.
[22:06:52] * johnny joined the chat.
[22:07:12] * mlundblad_netbook left the chat.
[22:07:28] * mlundblad_netbook joined the chat.
[22:11:12] * mlundblad_netbook left the chat.
[22:11:19] * mlundblad_netbook joined the chat.
[22:13:24] * slicer321 joined the chat.
[22:13:30] * slicer321 left the chat.
[22:15:07] * Guus joined the chat.
[22:35:12] * mlundblad_netbook left the chat.
[22:35:57] * alkino left the chat.
[22:36:42] * mlundblad_netbook joined the chat.
[22:38:12] * mlundblad_netbook left the chat.
[22:48:12] * stpeter left the chat.
[22:52:56] * luca tagliaferri left the chat.
[22:53:36] * luca tagliaferri joined the chat.
[23:14:12] * mlundblad left the chat.
[23:17:44] * johnny left the chat.
[23:28:20] * smoku left the chat.
[23:41:30] * johnny joined the chat.
[23:44:31] * hawke left the chat.
[00:00:51] * luca tagliaferri left the chat.
[00:08:39] * jugg joined the chat.
[00:24:45] * MattJ left the chat.
[00:27:43] * Guus left the chat.
[00:29:02] * Link Mauve left the chat.
[01:00:07] * evilotto left the chat.
[02:10:24] * Zash_ joined the chat.
[02:11:11] * Zash_ left the chat.
[02:12:22] * Zash_ joined the chat.
[02:12:28] * Zash_ left the chat.
[02:38:00] * johnny left the chat.
[02:40:48] * waqas left the chat.
[02:40:48] * waqas joined the chat.
[02:55:30] * lastsky joined the chat.
[02:58:45] * lastsky left the chat.
[03:01:40] * lastsky joined the chat.
[03:03:28] * scippio left the chat.
[03:03:30] * scippio joined the chat.
[03:26:29] * johnny joined the chat.
[03:33:57] * teo left the chat.
[03:34:52] * teo joined the chat.
[04:33:57] * Treebilou joined the chat.
[04:34:46] * Treebilou left the chat.
[04:34:48] * Treebilou joined the chat.
[04:48:36] * waqas left the chat.
[04:48:36] * waqas joined the chat.