Logs for jdev

Show join/part/nick changes:

[00:10:04] * Florob left the chat.
[00:30:39] * luca tagliaferri left the chat.
[01:03:43] * Zash joined the chat.
[01:04:56] * Lance joined the chat.
[01:05:29] * Zash left the chat.
[01:12:45] * nabatt joined the chat.
[01:28:05] * nabatt left the chat.
[01:49:43] * MattJ left the chat.
[01:56:00] * johnny left the chat.
[02:22:14] * johnny joined the chat.
[02:27:21] * Neustradamus joined the chat.
[03:08:24] * Lance left the chat.
[03:37:41] * Edward joined the chat.
[04:00:23] * Edward left the chat.
[04:07:20] * jkhii joined the chat.
[04:11:10] * Treebilou joined the chat.
[04:12:01] * Treebilou left the chat.
[04:12:29] * Treebilou joined the chat.
[04:41:02] * jkhii left the chat.
[05:07:31] * Treebilou left the chat.
[05:59:22] * mlundblad_laptop joined the chat.
[06:11:23] * Alex joined the chat.
[06:51:58] * Guus joined the chat.
[06:58:24] * bear left the chat.
[07:32:31] * teo1 left the chat.
[07:32:33] * teo1 joined the chat.
[07:59:55] * MattJ joined the chat.
[08:22:29] * luca tagliaferri joined the chat.
[08:29:33] * ermine joined the chat.
[08:31:23] * petermount joined the chat.
[08:51:10] * Tobias joined the chat.
[09:02:13] * smoku joined the chat.
[09:15:37] * nabatt joined the chat.
[09:23:32] * Tobias left the chat.
[09:41:37] * scippio_netbook joined the chat.
[09:45:24] * Treebilou joined the chat.
[09:46:06] * Treebilou left the chat.
[09:46:43] * Treebilou joined the chat.
[10:05:37] * alfeberlin left the chat.
[10:16:18] * Tobias joined the chat.
[10:22:35] * Tobias left the chat.
[10:23:30] * Tobias joined the chat.
[10:44:38] * scippio_netbook left the chat.
[10:44:54] * scippio_netbook joined the chat.
[11:00:50] * smoku left the chat.
[11:00:51] * smoku joined the chat.
[11:13:10] * smoku left the chat.
[11:13:11] * smoku joined the chat.
[11:26:32] * smoku left the chat.
[11:26:33] * smoku joined the chat.
[11:30:36] * Treebilou left the chat.
[11:57:00] * Tobias left the chat.
[11:57:06] * Guus left the chat.
[11:58:14] * Zash joined the chat.
[11:58:16] * Guus joined the chat.
[12:06:04] * Tobias joined the chat.
[12:12:36] * Guus left the chat.
[12:12:51] * Guus joined the chat.
[12:14:44] * bear joined the chat.
[12:20:30] * Zash left the chat.
[12:22:03] * tofu joined the chat.
[12:24:24] * smoku left the chat.
[12:27:54] * scippio_netbook left the chat.
[12:28:16] * smoku joined the chat.
[12:49:01] * Zash joined the chat.
[13:01:31] * Alex left the chat.
[13:16:49] * Tobias left the chat.
[13:54:57] * tong joined the chat.
[14:10:48] * lance joined the chat.
[14:10:48] * lance left the chat.
[14:48:57] * deryni joined the chat.
[15:05:07] * Guus left the chat.
[15:10:07] * mlundblad_laptop left the chat.
[15:22:07] * hawke joined the chat.
[15:22:43] * smoku left the chat.
[15:22:44] * smoku joined the chat.
[16:00:04] * luca tagliaferri left the chat.
[16:18:55] * nabatt left the chat.
[16:25:40] * Tobias joined the chat.
[16:26:30] * smoku left the chat.
[16:26:31] * smoku joined the chat.
[16:49:34] * teo1 left the chat.
[16:49:35] * teo1 joined the chat.
[16:49:38] * Tobias left the chat.
[16:55:10] * stpeter joined the chat.
[17:02:08] * stpeter left the chat.
[17:09:44] * petermount left the chat.
[17:11:56] * smoku left the chat.
[17:11:57] * smoku joined the chat.
[17:12:04] * stpeter joined the chat.
[17:16:13] * scippio left the chat.
[17:16:15] * scippio joined the chat.
[17:17:31] * scippio left the chat.
[17:18:31] <stpeter> heh
[17:19:06] * scippio joined the chat.
[17:19:08] <stpeter> johnny: my inbox is currently at ~3000, which is very far from zero :)
[17:20:24] <Zash> ^^
[17:33:24] * Tobias joined the chat.
[17:41:22] * smoku left the chat.
[17:46:06] * scippio left the chat.
[17:47:27] * scippio joined the chat.
[17:56:04] * scippio left the chat.
[18:03:24] * scippio joined the chat.
[18:12:45] * mlundblad joined the chat.
[18:15:28] * Treebilou joined the chat.
[18:33:41] * Asterix left the chat.
[18:33:58] * Asterix joined the chat.
[18:34:49] * scippio left the chat.
[18:49:10] * justin joined the chat.
[19:09:17] * lance joined the chat.
[19:09:17] * lance left the chat.
[19:10:36] * lance joined the chat.
[19:13:08] * Florob joined the chat.
[19:33:24] * hawke left the chat.
[19:33:39] * hawke joined the chat.
[19:36:00] * teo1 left the chat.
[19:36:01] * teo1 joined the chat.
[20:13:50] * lance left the chat.
[20:14:47] * smoku joined the chat.
[20:16:19] <justin> i'm interested to discuss persistent muc messages with nesting. this basically requires having ids on each message for future management of them (such as deletion) as well as indicating the thread hierarchy
[20:16:41] <justin> has there been any xep or effort regarding this?
[20:18:09] <louiz’> not as far as I know, just a recent little thread on the ML
[20:18:22] <justin> yeah that was my mail
[20:18:25] <justin> with no replies yet :p
[20:18:59] <johnny> lol
[20:19:07] <johnny> well regular messages have threads, why not mucs i guess
[20:19:17] <johnny> justin, the real problem is client implementation imo
[20:19:20] <johnny> that is.. after design
[20:19:34] <johnny> imo clients are what's holding back xmpp, because everybody has to "catch up"
[20:19:43] <johnny> google is doing the right thing with chrome auto updatesimo
[20:19:50] <johnny> we could be so lucky :(
[20:20:11] <justin> well this is for a custom client so i'm not real worried about that
[20:20:32] <justin> but i like to implement in a standard way if possible, so other clients may do the same in the future
[20:21:23] <Tobias> i've always found threading in mucs could be quite useful, if you have a decent interface to it
[20:21:38] <Tobias> especially when you have a lot people talking in a room
[20:23:03] <deryni> People have enough trouble getting email threads to work correctly when they need to explicitely reply to a message, I can't even imagine how clients would manage to automatically thread responses when users can just type responses.
[20:23:09] * edward joined the chat.
[20:23:50] <johnny> deryni, well there's a whole mess involved there
[20:23:56] <johnny> we have a little more freedom
[20:24:08] <johnny> the protocols are old, and implementations not so good
[20:24:09] <Kev> justin: Is there anything wrong with just using the message id of the parent as the thread id?
[20:24:31] <johnny> deryni, people seem to do ok with tweets tho..
[20:24:32] <Tobias> Kev, what message id?
[20:24:41] <Tobias> the id of the stanza?
[20:24:54] <Kev> Oh, right, I forgot that. I was going to suggest the room stamps a new message id on the stanza :)
[20:24:56] <johnny> it's a coarse interface , but it seems to work
[20:25:09] <deryni> johnny: Only sort of, and the reading interface for tweets is different than a standard muc room interface.
[20:25:16] <deryni> The possible "noise" is much higher, for example.
[20:25:31] <johnny> deryni, is it?
[20:25:37] <johnny> how so?
[20:25:41] <johnny> that's how identichat worked
[20:25:45] <Tobias> Kev, since IIRC one can only expect those ids to be unique for a single xmpp stream session, and not globally unique or so
[20:25:47] <johnny> a muc interface to microblogs
[20:26:01] <Kev> Tobias: But the MUC can enforce uniqueness of them.
[20:26:01] <deryni> Yes, but that's not the standard reading interface.
[20:26:09] <justin> right it would be ids unique per room
[20:26:12] <johnny> so, you can do the same thing
[20:26:21] <Tobias> Kev, right, if the MUC would stamp them on it'd work
[20:26:50] <justin> i'm still not sure if stanza id is appropriate for that, since then it's not new for every delivery, and might be weird with tracking error stanzas
[20:27:03] <justin> but i do like reusing it if possible
[20:27:19] <deryni> johnny: You're fighting the wrong side of my argument. I'm saying you need to come up with an interface that makes people feel the *need* to do direct replies to specific messages (which actually doesn't happen in twitter as often as a simple targetted reply, as I understand it).
[20:27:55] <johnny> ah yes, the *need* that's what twitter did. that we could not
[20:28:05] <deryni> A client could attempt to be smart about auto-threading any message with a "<nick>:" prefix but that seems as likely to be wrong as to be right.
[20:28:07] <johnny> i mean.. not just in xmpp, but in general
[20:28:15] <johnny> deryni, it only has to be "good enough"
[20:28:24] <deryni> Not to mention that, as often as not, people don't direct every message to the person they are talking to.
[20:28:24] <johnny> twitter sucks
[20:28:25] * ermine left the chat.
[20:28:28] <johnny> but it still works
[20:28:30] <justin> so the simple idea is: user submits comment without message id (or arbitrary id), and can optionally pass <thread parent="parent_id"/> (no text node!). muc assigns id to message and echos out
[20:29:03] <johnny> deryni, our problem is probably perfection.. twitter proves that it doesn't always matter :)
[20:29:13] <justin> s/comment/message (you can see my application terms trickling into the discussion...)
[20:29:54] * smoku left the chat.
[20:29:58] * smoku joined the chat.
[20:30:05] <justin> deryni: in our interface, you select exactly what message you are replying to
[20:30:23] <deryni> Right, in a custom interface the replying issue can be made to go away.
[20:30:35] <deryni> I'm talking about generic xmpp/muc clients.
[20:30:59] <justin> hmm.. right
[20:31:18] <deryni> Backing up a web forum via a much room with threads would be entirely doable (for example), though I'm not sure why one would want to go that route.
[20:31:23] <deryni> s/much/muc/
[20:31:33] <MattJ> justin, I read your mail, but haven't replied yet
[20:31:48] <justin> deryni: heh ;)
[20:31:50] <MattJ> justin, I'm not sure I see the issue, and I'm not sure you would create a unique threadid for each message
[20:31:55] <deryni> The problem is when you need to have both "direct" and "normal" messages from the same interface.
[20:32:46] <justin> MattJ: each message does need some sort of unique id though, so that a reply can be targetted to a specific parent message
[20:33:48] <MattJ> Yeah, so that's kind of an in-reply-to, as well as starting a new thread
[20:34:14] <MattJ> ok, so you do need some per-message unique id :)
[20:34:15] <deryni> You want/need branching threads? (I haven't read the email yet.)
[20:34:41] <Tobias> one could also have an implicit id based on a hash of the message, but that feels a bit like taking a sledgehammer to crack a nut
[20:34:54] <MattJ> deryni, <thread> has a 'parent' attribute, but that specifies a parent thread, not a parent message
[20:35:09] <MattJ> deryni, and a typical thread tree of messages needs that
[20:35:53] <justin> yes, id + in-reply-to is what is needed (and i considered SHIM for this, though the ids are based on those weird email ids)
[20:36:01] <deryni> That was my question though. If you only need an id to tie a message to a thread (and let date/etc. handle the sorting from there) or whether you need to be able to hand messages off of other individual messages.
[20:36:15] <deryni> s/hand/hang/
[20:36:16] <MattJ> The latter
[20:36:34] <deryni> Ok, that's what I figured.
[20:36:44] <MattJ> Someone could reply to a message after someone already replied to a later message
[20:37:35] <Tobias> right, this already happens quite often in email discussions :)
[20:38:14] * Tobias left the chat.
[20:38:22] <deryni> Right, I wasn't questioning why someone would need that just whether that was what was being specifically discussed.
[20:38:28] * Tobias joined the chat.
[20:43:34] <justin> MattJ: so are you thinking there should be a different way to specify in-reply-to rather than using thread?
[20:44:03] <MattJ> That's what I'm leaning towards, but this is possibly unchartered territory (which may be why no-one has replied on the ML)
[20:44:38] <Kev> MattJ: What was the problem with using thread in the way I suggested?
[20:44:54] <MattJ> stanza ids aren't guaranteed to be unique
[20:45:02] <Kev> They are if they're created by the MUC.
[20:45:10] <MattJ> Are they?
[20:45:14] <justin> do threads have to be "unique" ?
[20:45:19] <MattJ> They do
[20:45:24] <Kev> Sure, if the MUC creates them, it can have any rules it wants.
[20:45:29] <MattJ> Well, distinct threads do :)
[20:45:42] <MattJ> The XEP recommends a UUID
[20:45:48] <MattJ> but you could use a URL, or something
[20:45:48] <justin> i just mean, there isn't anything in rfc3921 that says thread ids have to be globally unique
[20:45:56] <MattJ> !xep 201
[20:45:57] <Kanchil> MattJ: XEP-0201: Best Practices for Message Threads is Informational (Proposed, 2010-09-28) See: http://xmpp.org/extensions/xep-0201.html
[20:46:02] <MattJ> and there might be in -bis
[20:46:09] <Kev> bis references 201
[20:46:16] <MattJ> I'm not sure if it was added there or moved o... right :)
[20:46:37] <Kev> Doesn't it?
[20:47:40] <Florob> Yes, bis tells you to look at 201 for more details. When actually I had to look at bis to find the details I had anticipated in 201, but that's another story
[20:51:57] <stpeter> wow, I missed lots of messages here
[20:52:22] <stpeter> we removed the UUID restriction
[20:52:33] <stpeter> do threads really need to be globally unique?
[20:52:45] <Kev> No, as we discussed onlist :)
[20:52:51] <stpeter> didn't think so :)
[20:52:52] <stpeter> brb
[20:52:54] <Tobias> not globally, but room wide
[20:53:04] <Kev> But for Justin's application he may choose to make them unique room-wide :)
[20:54:05] <justin> ok, so the question now is whether using a different thread for every message counts as an abuse of threads
[20:54:23] <MattJ> By now I'm inclined to say no, just go and get coding :)
[20:54:39] * Neustradamus left the chat.
[20:54:40] <Kev> Well, that would be silly, no?
[20:54:48] <Kev> Ah, right, because of the parent id?
[20:54:52] <MattJ> Right
[20:54:55] <Kev> Yeah, that's horrible but it works fine.
[20:55:06] <MattJ> Necessarily horrible
[20:55:07] <Kev> I don't see a reason for anything more complicated.
[20:56:06] <Florob> well, I personally would certainly call it abuse. It's not like it's really bad, but if we had a need for it outside a closed environment I'd rather we have a pretty much equivalent second XEP that's not threads, but message replies
[20:56:28] <Kev> Florob: I don't see any harm in this.
[20:56:29] <Florob> because you might want both at the same time
[20:56:38] <Kev> Why? One supercedes the other.
[20:58:02] <Florob> doesn't one basically require one thread ID per message and the other same thread IDs among multiple messages?
[20:58:14] <Kev> What I mean is:
[20:58:34] <Kev> If you have a need for thread hierachies, you are automatically granted flat threading as well.
[20:58:38] <Kev> It's just a richer model.
[21:00:21] <justin> but if you have a tool that visualizes threads based on what the standard says (first message and all replies have the same thread), then this would probably fail to visualize properly on the system we're discussing
[21:00:42] <Florob> Well, as I said, in a constrained environment, yes. But as soon as a "normal" thread exists, it becomes impossible to reply to a specific message.
[21:01:12] <stpeter> /me scrolls up again
[21:02:39] <Kev> Florob: I take the point.
[21:02:53] <stpeter> ah, good, the one document I hadn't yet reviewed for tomorrow's IESG telechat has been deferred by someone else :)
[21:03:09] <stpeter> I love it when that happens :)
[21:04:19] <justin> the overkill approach is to have a way to specify the message id being replied to. so you have root message (A1) and five replies (B1-5), all sharing the same <thread> value. then someone replies to B3 to create C1, specifies <thread parent="first">second</thread> and *additionally* indicates some other way that this is a reply specifically to B3 and not just to the first thread (<in-reply-to>B3</in-reply-to> or such)
[21:07:09] * mlundblad left the chat.
[21:14:19] * Neustradamus joined the chat.
[21:18:01] * hawke left the chat.
[21:18:38] * hawke joined the chat.
[21:42:40] * stpeter left the chat.
[21:43:32] * Tobias left the chat.
[22:00:08] * scippio joined the chat.
[22:01:44] * jcea joined the chat.
[22:21:16] * johnny left the chat.
[22:31:07] * stpeter joined the chat.
[22:33:56] * johnny joined the chat.
[22:50:54] * louiz’ left the chat.
[22:51:09] * louiz’ joined the chat.
[22:53:25] * Zash left the chat.
[22:58:24] * smoku left the chat.
[22:58:48] * smoku joined the chat.
[22:59:28] * smoku left the chat.
[22:59:43] * smoku joined the chat.
[23:05:53] * hawke left the chat.
[23:06:20] * hawke joined the chat.
[23:11:47] * louiz’ left the chat.
[23:11:51] * louiz’ joined the chat.
[23:18:04] * tong left the chat.
[23:18:58] * MattJ left the chat.
[23:27:50] * Florob left the chat.
[23:46:38] * hawke left the chat.