Home
jdev
jdev@conference.jabber.org
Sunday, April 1, 2007< ^ >
Alex has set the subject to: Jabber/XMPP development (english spoken room)
Room Configuration

GMT-5
[00:21:30] jeeterm joins the room
[01:15:07] sjr joins the room
[01:51:07] sjr leaves the room
[01:57:15] bct leaves the room
[02:01:23] jeti joins the room
[02:31:52] TobiasFar joins the room
[02:54:47] elmex joins the room
[02:54:59] <elmex> did someone miss me?
[02:55:23] <TobiasFar> not me ;)
[02:56:22] <elmex> :-(
[02:56:24] <elmex> you are mean
[02:57:11] <TobiasFar> not really
[02:57:19] <elmex> i thought i've found friends for the first time in my life and you... you... you dont like me *sob*
[02:57:43] <TobiasFar> heh...seems you have to life with it
[02:58:16] <TobiasFar> but another question: How does the perfect roster have to look like? What features should it have?
[02:59:50] <elmex> the perfect roster should show me whether the other person is online and since when. it should also be space efficent. meaning: show most information with a good info/space ratio
[03:00:35] <TobiasFar> but if you have a good info/space ratio the font will be pretty small
[03:00:45] <elmex> i don't want icons which are 2 times bigger than the font just to show me that he is online. its enough that he shows up. (meaning: offlien users aren't shown if i don't want them to)
[03:01:05] <elmex> well, the font should be user-settable
[03:01:28] <elmex> let him pick the one that best suits him. but don't take up so much extra-space
[03:01:34] <TobiasFar> what's with avatars?
[03:02:32] <elmex> avatars are... colorful. but i don't need them for the usage and they usually don't represent information that i need to see on the roster.
[03:03:01] <elmex> but i'm not the average windows-user-wants-colorful-and-unuseful-programs-user
[03:03:18] <elmex> that might be too a preferecence
[03:03:37] <TobiasFar> what's about structure? just a list or tree structued
[03:04:07] <TobiasFar> or maybe something entirely different
[03:04:08] <elmex> how many levels has the tree?
[03:04:43] <TobiasFar> depends on the usage of groups and meta contacts and whether resources are shown
[03:05:54] <lynx> hmm.. my roster is a text line on top of my telnet :)
[03:06:16] <elmex> well. some kind of tree view is ... nice... but i've just had a crazy idea: it would maybe be useful if the most 'used' contacts would show up at the top. i mean that the roster sorts itself by counting the frequency with wich i contact someone.
[03:07:03] <elmex> but aside from that: i guess a list structured in groups will be best.
[03:08:12] <TobiasFar> hmmm...
[03:08:18] <TobiasFar> thanks for your inpuit :P
[03:08:49] <elmex> the most important thing if you want broad user acceptance: a good default and an easy way to make preferences.
[03:09:37] <elmex> icons and avatars are 'nice' in the roster. i guess i would like to view them sometimes too. but not all the time
[03:10:04] <elmex> it's enough for me that offline users or away users are shaded or something like that
[03:10:18] <elmex> (offline users best not shown at all)
[03:10:39] <TobiasFar> yeah...nice idea...to fade them away the more inpresence they are
[03:24:56] <elmex> the important thing is that it should allow me to reach the one i want to reach as quick as possible, even if i have a big roster
[03:26:27] <fippo> sort users by the time you've talked to them last (or use a self-ordering list)
[03:26:47] <elmex> why talked to 'last' ?
[03:27:07] <elmex> i would rather like it to be sorted by contact frequency
[03:27:20] <elmex> because the propability that i want to reach them is higher
[03:27:28] <TobiasFar> yeah...contact frequency sounds better to me too
[03:27:35] <lynx> and then have to look for people in a mess? i'm pro for groups
[03:27:43] <fippo> which is what i mean by "by the time you've talked to them last"
[03:28:16] <lynx> when you're looking for someone you dont call often, she might be anywhere in the list
[03:28:18] <elmex> lynx: well, group view should still be possible. but for me it shouldn't be the regular one
[03:28:29] <elmex> lynx: for that you need a good search function
[03:28:46] <lynx> i dont even care to have group seperators.. just order by groups
[03:28:50] <fippo> probably create an artificial "frequent contacts" group
[03:30:45] <elmex> well,groups is something i've to 'administrate' myself, i've to take care they all end up in the right group when subscribing... i don't want to spend time to adminstrate such an artificial seperation. it's not bad to have, but it's not that important to me. maybe even add a mode to sort by most-frequent-group or something like that
[03:31:32] <elmex> roster are suboptimal anyway
[03:31:58] <elmex> i don't want to see everyone who is online except when i really want to see them
[03:32:21] <elmex> there should be no window at all :->
[03:32:35] <elmex> (but i manage that with my window manager)
[03:32:38] <TobiasFar> how does the program know when do you want to see somebody
[03:32:50] <elmex> by telling the program :->
[03:32:58] <TobiasFar> it already knows that mostly you don't want to see the offline ones
[03:32:59] <elmex> (click on it or press a shortcut)
[03:33:05] <elmex> right
[03:33:23] <elmex> but i don't want to see all online users too... except maybe some special ones (like my girlfriend)
[03:33:37] <elmex> but that can be managed somehow by groups
[03:33:40] <elmex> but that suboptimal
[03:33:57] <TobiasFar> why do you subscribe presence from people you don't want to see online?
[03:34:15] <elmex> my gf is in the 'family' group... and thats where she belongs. but putting her into a 'special' group i always display hides that relation
[03:35:20] <elmex> TobiasFar: because sometimes i like to contact them, and then it's most easiest to see whether they are online (whether talking to them will be most likely successful)
[03:35:20] <fippo> tobiasfar: because unsubscribing them is considered impolite (and you therefore do "polite blocking")
[03:35:39] <lynx> i tend to have two major groups too.. the ones i need to talk to, or might need to.. and the ones i only lookup occasionally..
[03:35:46] <TobiasFar> lol
[03:35:51] <lynx> so the first group is visible, the second one is packed away
[03:36:06] <elmex> yes, but you need to administrate that stuff yourself
[03:36:25] <lynx> ideally i would like to only get alerts from the visible group (logged in, logged out etc)
[03:36:25] <elmex> i don't want to do constant group-changing..
[03:36:49] <elmex> yes, such alerts on request would be nice
[03:36:56] <TobiasFar> elmex: yeah..you should have to define the group only at subscribing time
[03:36:57] <elmex> 'inform me when my gf comes online'
[03:37:10] <elmex> TobiasFar: right
[03:37:14] <elmex> and sometimes not even then
[03:37:18] <lynx> my telnet interface is totally alert-oriented.. one line of text shows me who's coming or going
[03:37:30] <elmex> right
[03:37:33] <lynx> but i need to filter it to the people that are information to me
[03:37:37] <lynx> or otherwise presence = noise
[03:37:50] TobiasFar likes the big and fancy popup way
[03:38:02] <elmex> i would love a im-clien that runs in a terminal window like that too (thats maybe thereason why i am writing such a client)
[03:38:04] <lynx> the buddy list as such is usually not so important, because i already read who's there
[03:38:51] <lynx> oh no.. popups that break into my work, never.. i like all events nicely lined up in my chat window..
[03:38:52] <elmex> TobiasFar: popups are too distracting for me. when i'm on another screen and i'm coding, then the last thing i want is a window popping up and distracting me from the code
[03:39:29] <lynx> whatever it is.. changes to the wiki, emails, privmsgs, syslogd panicking
[03:39:33] <elmex> TobiasFar: but there also hits the 'make preferences easy' rule :)
[03:39:54] <lynx> headline news.. it's all in one window that i periodically read & clear
[03:40:24] <lynx> i even monitor access to my websites that way
[03:41:02] <lynx> did i mention it's easier to do psyc gateways in small apps than xmpp? ;)
[03:41:26] <fippo> and you frequently talk in the wrong room because you only have one window for all :-)
[03:41:51] <elmex> nope
[03:41:58] <elmex> ah, right, i don't do that anymore
[03:41:58] <lynx> YES that's the stupidest problem about single-window chat interfaces
[03:42:10] <elmex> i solved that problem
[03:42:22] <elmex> i will never write in the wrong one again
[03:42:29] <elmex> and still have a curses interface
[03:42:30] <TobiasFar> what's the problem with single-chat windows?
[03:42:33] <lynx> but that's just because i also have all chatrooms in the same window, which isn't a recommendable approach and i'm willing to upgrade to a decent psyc client
[03:42:50] <elmex> TobiasFar: the main problem was with irssi: multiple chat windows but only one input line
[03:43:01] <TobiasFar> ahh
[03:43:04] <elmex> i happened to have the wrong window focused quite often
[03:43:24] TobiasFar doesn't use CLI programs that often
[03:43:25] <elmex> like writing in the public: "you are so hot, i want to stick it into your ******"
[03:43:40] lynx asks: psycion uses a screen(1)-like interface.. what does bummskraut use?
[03:43:40] <elmex> when i wanted to write that to my gf
[03:44:22] <elmex> lynx: bummskraut has multiple buffers. and you can connect multiple frontends. and each frontend can have a different buffer selected
[03:44:38] <elmex> a buffer is a 'contact' of some kind
[03:45:40] <lynx> for the girls a client with good love action visualisation would be great
[03:45:53] <elmex> ?
[03:46:10] <elmex> you mean with a window that shows a porno to make them bed-ready?
[03:46:22] <lynx> like i type /kiss and she gets a boom bang fantastic rhps-like animation all over her screen
[03:46:22] xia212000 joins the room
[03:46:30] <elmex> haha
[03:47:02] lynx asks: so elmex, you use multiple processes in a screen or xterms?
[03:47:14] <lynx> reminds me of /window create of ircII
[03:47:16] <elmex> in rxvts :)
[03:47:38] <elmex> or in screens
[03:47:41] <elmex> whatever you like
[03:47:46] <lynx> ircII has the ability to open screen(1)'s or xterms.. hardly anyone figures that out
[03:48:03] <lynx> it's in the code since 91 or something
[03:48:06] <elmex> i guess it's something like that
[03:48:26] <elmex> where you start the frontend process (in screen or somewhere else) doesn't matter
[03:48:33] <elmex> (you can even run it on a different machine)
[03:48:42] <elmex> (like i usually do)
[03:48:51] <lynx> yeah yeah your bouncer protocol is pretty neat, elmy ;)
[03:49:01] <elmex> well, it's not prefect
[03:49:05] <lynx> to bad it's not psyc.. why json? *g*
[03:49:09] <elmex> but it solves my problems
[03:49:49] <lynx> (not a serious question, json is pragmatic for single source apps)
[03:49:50] <elmex> json was easy to parse as i already had a parser for that, that is easily pluggable and has a very defined interface
[03:50:09] <elmex> to_json and from_json. thats all :->
[03:50:39] <fippo> elmex: the same is true for macromedias AMF :-P
[03:51:06] <elmex> whats amf?
[03:51:08] <lynx> so you totally and convincedly exclude the options of (a) having differing versions on various machines and (b) having more than one codebase for the protocol :)
[03:51:49] tomprince leaves the room
[03:51:49] tomprince joins the room
[03:51:50] tomprince leaves the room
[03:52:15] <elmex> lynx: well, the frontends can of course be in different versions... once the protocol is stable enough it can be implemented by anyone.
[03:52:19] tomprince joins the room
[03:52:31] <elmex> lynx: and why can't i have multiple code-bases?
[03:52:39] <lynx> i disagree - either you can just use the json data, then you cant have differing versions
[03:52:51] lynx asks: or you are implementing a protocol on top of json, then why a new protocol?
[03:52:51] <TobiasFar> someone familiar with FastCGI?
[03:53:25] <elmex> lynx: json is just the 'encoding' of data structures. the data structures are of course meaningful and have a protocol.
[03:53:39] <elmex> s/have/follow/
[03:53:49] <elmex> lynx: a new protocol that suits the job best
[03:54:09] <elmex> it's better than abusing irc for bouncing in the manner i needed
[03:54:17] <lynx> so even if json gives things as is, you still have to check their meaning and sanity
[03:54:26] <elmex> of course
[03:54:42] <elmex> like in any protcol you still have to apply semanticcs somewhere
[03:54:59] <lynx> ok, better than abusing irc is always good.. but extensible protocols like jabber or psyc should give you a hand here and there
[03:55:57] <lynx> ok xml probably gives you more trouble than it actually helps you ;) ;)
[03:56:14] <elmex> (jabber isn't extensible, but others might differ. XEPs are just protocol hacks from my point of view. adding a 'new' element somewhere where it is usually ignored by non-supporting clients is easy)
[03:56:40] <elmex> with that even IRC is extensible
[03:57:25] <elmex> lynx: the problem with psyc was that i didn't have a parser for the protocol and no defined interface.
[03:57:34] <lynx> append at the end of a line and hope some other irc dialect doesn'T append something completely different... brrrr
[03:57:44] <elmex> :)
[03:57:52] lynx asks: which language, elm?
[03:58:00] <lynx> ah right.. perl..
[03:58:03] <elmex> lynx: perl :)
[03:58:09] <lynx> hey.. we have Net::PSYC since 1999!!
[03:58:58] lynx asks: and since you want to do your own methods anyway.. why should the flowing state of the psyc unstandard disturb you?
[03:59:00] <elmex> Net::PSYC is great, but it didn't offer me a interface for just protocol parsing. and it took me too much power to customize it. and most important: it doesn't work so good anymore - at least as far as el told me
[03:59:37] <elmex> (at least wrt to real world usage)
[04:00:07] <elmex> json was there, json worked, i didn't have to write my own parser, so i used it.
[04:00:21] <elmex> switching the protocol to psyc is still possible
[04:00:29] <lynx> eh? then just revert to my version before el starting coding on it!! hahaha kidding.. in fact you can always just receive and send msgs with a rich interface.. no matter which methods you use.. you don't have to use the whole Net::PSYC::Client logic
[04:00:37] <elmex> most of the protocol handling is abstracted away pretty nicely
[04:01:08] <lynx> i run Net::PSYC every time i fire up my psycmp3 player.. no problems at all
[04:01:28] <elmex> heh, ok
[04:02:44] xia212000 leaves the room: Disconnected
[05:09:23] smoku joins the room
[05:15:13] queltos joins the room
[05:47:06] Kev leaves the room
[05:47:14] Kev joins the room
[05:53:35] jeti leaves the room
[05:55:27] elmex leaves the room
[06:01:31] legoscia joins the room
[06:04:39] smoku leaves the room
[06:26:53] TobiasFar leaves the room
[06:41:55] TobiasFar joins the room
[06:56:26] niekie leaves the room
[07:14:39] queltos leaves the room
[07:45:52] TobiasFar leaves the room: Replaced by new connection
[07:45:57] TobiasFar joins the room
[08:03:35] Solarius joins the room
[09:31:09] jeeterm leaves the room
[09:31:09] lynx leaves the room
[09:31:31] lynx joins the room
[09:33:32] <lynx> anybody please ask my vCard.. i'd like to see if i successfully utf8'ed my profile data......
[09:34:42] <lynx> yeah! no xml crash this time :)
[09:36:58] jeti joins the room
[09:38:33] <legoscia> looks right to me
[09:40:06] lynx asks: does your server tell you when somebody looks up your vcard?
[09:41:52] <TobiasFar> on direct requests my client can tell me
[09:42:51] <lynx> i'm undecided whether to make it show or not.. ah yeah.. in xmpp it takes a xep for that hehe.. i just make up a _notice for it on the fly :)
[09:56:25] fippo leaves the room
[10:00:38] fippo joins the room
[10:31:44] jeeterm joins the room
[10:31:49] jeeterm leaves the room
[10:32:00] jeeterm joins the room
[10:33:32] jeeterm leaves the room
[11:34:25] legoscia leaves the room
[11:51:27] bct joins the room
[12:05:10] njoyce joins the room
[12:06:34] <njoyce> If a client makes a connection to a XMPP server, and for some reason the client was not allowed to connect, how would the server say that in XMPP and then close the connection?
[12:06:35] iono leaves the room
[12:42:09] <fippo> njoyce: policy-violation stream error seems most appropriate to me
[12:44:15] <njoyce> fippo: looks good to me, thanks!
[12:52:24] niekie joins the room
[12:52:44] njoyce leaves the room
[13:12:25] tomprince leaves the room: Replaced by new connection
[13:12:25] tomprince joins the room
[13:12:28] tomprince leaves the room
[13:12:43] tomprince joins the room
[13:13:44] tomprince leaves the room: Replaced by new connection
[13:13:45] tomprince joins the room
[13:13:46] tomprince leaves the room: Replaced by new connection
[13:13:48] tomprince joins the room
[13:13:50] tomprince leaves the room
[13:13:56] tomprince joins the room
[13:21:02] sjr joins the room
[13:49:42] sjr leaves the room
[13:53:07] legoscia joins the room
[14:05:07] tomprince leaves the room
[14:05:26] tomprince joins the room
[14:05:36] tomprince leaves the room
[14:06:30] tomprince joins the room
[14:23:30] sjr joins the room
[15:01:28] jeti leaves the room
[15:08:48] tomprince leaves the room
[15:13:46] jeti joins the room
[15:21:52] jeti leaves the room
[15:22:39] Solarius leaves the room
[15:26:54] tomprince joins the room
[15:41:59] jeti joins the room
[15:43:30] sjr leaves the room
[15:53:08] Kenck joins the room
[15:53:45] Kenck leaves the room
[15:54:41] Alex joins the room
[15:54:41] Alex leaves the room
[15:59:43] Kenck joins the room
[16:00:13] <Kenck> any girl for me
[16:00:39] <Kenck> hello
[16:01:03] <Kenck> /legoscia: hello
[16:02:15] <TobiasFar> Kenck: wrong room
[16:02:43] <Kenck> Tobias > because
[16:03:08] <TobiasFar> topic
[16:03:37] Kenck leaves the room
[16:13:02] <legoscia> btw, the right
[16:13:09] <legoscia> btw, the right room is chat@conference.jabber.org ☺
[16:17:41] jeti leaves the room
[16:51:20] TobiasFar leaves the room
[16:53:52] legoscia leaves the room
[17:19:34] maddrick joins the room
[17:24:10] maddrick leaves the room
[17:34:36] sjr joins the room
[18:17:34] bct leaves the room
[18:18:26] bct joins the room
[18:37:42] sjr leaves the room
[18:42:55] Yandris joins the room
[18:43:11] Yandris is now known as Jost
[18:43:34] <Jost> Hello
[18:45:11] Jost leaves the room
[18:47:59] Jost joins the room
[18:48:20] <Jost> Hello
[18:48:45] Jost leaves the room
[20:26:50] sjr joins the room
[21:21:23] NICKY_JAM joins the room
[21:23:39] <NICKY_JAM> que bola
[21:23:40] tomprince leaves the room
[21:25:51] <NICKY_JAM> pinga habln cojoneeeeeeee
[21:26:30] NICKY_JAM leaves the room
[22:49:05] jeeterm joins the room
[22:51:02] sjr leaves the room
[23:15:51] sjr joins the room
[23:25:59] niekie leaves the room
[23:47:48] bct leaves the room
[23:49:11] sjr leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!