User:Rob Linden/Office Hours/2007 Jun 15

From Second Life Wiki
Jump to: navigation, search
[15:01] Stevex Janus: Hi rob
[15:01] Rob Linden: hi
[15:01] Dale Glass: hiya :-)
[15:01] Rob Linden gets used to new interface for text chat
[15:02] Dale Glass: what new interface?
[15:02] Stevex Janus: what new interface :-
[15:02] Rob Linden: oh, this might only be in the First Look viewer
[15:03] Rob Linden: the way for dealing with chat history is a little different
[15:04] Stevex Janus: I'm still not clearn on the difference between First look and the source releases.
[15:05] Ruud Lathrop: wow nice office
[15:05] tessa Amos: :)
[15:05] Rob Linden: Stevex: it varies
[15:06] Rob Linden: there are generally three types of binary drops we do, which we generally do source releases for
[15:06] Rob Linden: 1. Main grid releases
[15:07] linux soda machine Mark II v1.05: Machine is in use by tessa Amos
[15:07] Rob Linden: 2. First Look (which is a pre-release viewer against the main grid)
[15:07] linux soda machine Mark II v1.05: Here's your can of Ubuntu flavored Linux soda, enjoy!
[15:07] Rob Linden: 3. Beta (which is pre-release viewer against pre-release grid)
[15:07] Rob Linden: We often publish source whenever we publish a binary
[15:07] Ruud Lathrop: So how is unbuntu baby, tast good?
[15:07] tessa Amos: lolol
[15:08] Stevex Janus: Ok I see
[15:08] tessa Amos: kinda tastes like diet pepsi :P
[15:08] Ruud Lathrop: Hmmm dont like pepsi
[15:08] Rob Linden: we're making an exception this go around, because there's still a few things that we need to tease out of the current First Look that we don't have the right to release in source form
[15:08] Stevex Janus: so the stuff that I'm compiling is for the main line.
[15:08] linux soda machine Mark II v1.05: Here's your can of Red Hat Fedora flavored Linux soda, enjoy!
[15:08] Rob Linden: Stevex: probably.
[15:08] Stevex Janus: yea.
[15:08] Rob Linden: yup, that's the main release
[15:09] Stevex Janus: I've been trying to collect profileing info on the client
[15:09] Stevex Janus: Lots of time spent in the skycalc
[15:10] Rob Linden: Dale, you should take a look at the First Look release, actually. The new "near me" tab is sorta like your a/v scanner, though not really
[15:10] Dale Glass: hmm, will do
[15:10] Rob Linden: your stuff is a lot more specialized
[15:11] Dale Glass: btw, any chance you could do source releases a bit faster? The last SVN release took quite a while after the grid came back up
[15:11] Rob Linden: the tarball came out pretty quickly, but I wasn't able to get to teh svn release until later
[15:11] Rob Linden: I'm still doing that stuff personally, and it's not fully automated yet
[15:12] Dale Glass: yeah, I'm moving towards SVN. Used to use tarballs, but that's quite a pain to manage. Now I'm moving over to svk
[15:13] Dale Glass: hiya :-)
[15:13] Saijanai Kuhn: yo
[15:13] Stevex Janus: hello
[15:13] Joshua Linden: Oops, didn't mean to drop in so dramatically.
[15:13] Rob Linden: hi folks
[15:13] Saijanai Kuhn: was very dramatic I'm sue
[15:13] Saijanai Kuhn: *sure
[15:13] Rob Linden: I was just putting out word on the #opensl channel that this was going on
[15:13] Basic Chair: Right click me and choose 'Sit Here' to sit down
[15:14] Rob Linden: before I get done typing, three people show up, including a Linden :)
[15:14] Basic Chair: Right click me and choose 'Sit Here' to sit down
[15:14] Wyn Galbraith: Hey Joshua, a new Linden!
[15:14] Rob Linden: actually, Joshua predates me (by a couple of weeks, I think)
[15:14] Wyn Galbraith notes she collects Linden bears *winkwink*
[15:14] Joshua Linden: I'm just here to listen, don't mind me.
[15:15] Wyn Galbraith: I love the pirate.
[15:15] Joshua Linden: Yeah, my rez day is coming up in a week or so.
[15:15] Stevex Janus: Rob, would it be possible to include a gcc4.1.1 libiary for the KDU stuff?
[15:15] Wyn Galbraith: Is that like a birthday?
[15:16] Rob Linden: I'm not sure. We might be able to do a one-off build. which distros have gcc4.1.1 these days?
[15:16] Stevex Janus: FC6
[15:17] Gigs Taggart: hi
[15:17] Stevex Janus: I don't know about others.
[15:17] Rob Linden: I guess Red Hat always likes to stay on the bleeding edge w/compilers
[15:17] Gigs Taggart: heh not always, they used that broken 2.96 for ages
[15:17] Dale Glass: gentoo has gcc4 too
[15:18] Rob Linden: I can't recall what we compile with by default.
[15:18] Dale Glass: although gentoo is flexible on that regard
[15:18] Stevex Janus: I was going to compare the diff between the gcc3.x and gcc4.x on the client.
[15:18] Gigs Taggart: 3 is default
[15:18] Gigs Taggart: 4 needs minor patches
[15:18] Dale Glass: SConstruct specifies 3.4
[15:18] Gigs Taggart: yeah
[15:18] Rob Linden: ah...ok
[15:19] Stevex Janus: I have a patch to alllow you to set gcc version. I will submit that to jira
[15:19] Rob Linden: oh, cool
[15:19] Saijanai Kuhn: mac ox x 4+ looks like it comes with Gcc4.x
[15:19] Rob Linden: I take it gcc 4 mucks with the abi again?
[15:19] Dale Glass: that is good :-) I'll vote on that
[15:19] Stevex Janus: Yea, there was a change in 4.1
[15:20] Gigs Taggart: I think you all should consdier doing 4 for your build, it's very wide spread now
[15:20] Rob Linden: 4.0 and 4.1 have different abi?
[15:20] Stevex Janus: I did but but it is dog slooooowwww
[15:20] Gigs Taggart: slow eh... did they change the optimization that much?
[15:21] Stevex Janus: I would have to go back ver the gccmailing list but I do recall there was a C++ ABI change
[15:21] Gigs Taggart: 4 should have more autovectorization I would think
[15:21] Gigs Taggart: don't know if that's made it out of beta
[15:21] Stevex Janus: The client is slow because when using gcc4 the KDU libs don't link in. I thinhk
[15:22] Gigs Taggart: ah yeah
[15:22] Rob Linden: Stevex: could you file a request in jira for gcc 4.x/Linux KDU build?
[15:22] Rob Linden: (or someone)
[15:22] Stevex Janus: Ok
[15:22] Rob Linden: (if it isn't already filed)
[15:22] Dale Glass coughs
[15:22] Dale Glass: damn, they set oil on fire on the kitchen or something O.o
[15:22] Gigs Taggart: maybe you'll get steaks for dinner :)
[15:23] Gigs Taggart: Rob is there an agenda for this meeting?
[15:23] Rob Linden: sadly, not a predetermined one
[15:23] Gigs Taggart: heh, that's not sad really :)
[15:23] Gigs Taggart: I want to talk about triage policy and Dzonates
[15:23] Rob Linden: sure...go for it
[15:23] Rob Linden: I was just about to ask
[15:23] Gigs Taggart: I don't understand why he has tabled all the bugs on the agenda because of "lack of votes"
[15:24] Gigs Taggart: is there some new system I don't know about?
[15:24] Stevex Janus: Rob, no KDU issues. so will make one
[15:24] Rob Linden: no...he's sorta acting on his own, though using votes as a filter isn't an altogether bad idea
[15:24] Gigs Taggart: I think it is, when there's a clear repro, it's something that should be fixed, but it's not somethign that the user on the street would understand or care about
[15:25] Gigs Taggart: I thought the whole point of triage was to get that sort of thing looked at, but if we only look at votes those will never get hit
[15:26] Rob Linden: I wouldn't use that as the be-all, end-all metric, but anything that meets your description should be able to muster a couple of votes
[15:27] Rob Linden: I'd want to set the bar pretty low on tabling by lack of votes; but >1 isn't too much to ask. what is the bar that Dz is using?
[15:27] Gigs Taggart: I kinda liked just skimming high voted bugs and putting them in each triage
[15:27] Gigs Taggart: and then going sequentially otherwise
[15:28] Fremont Cunningham: SL is so full of 'gaming', using votes that can be stuffed is risky.
[15:28] Gigs Taggart: I guess the fundamental question is "what is the goal of triage?"
[15:28] Gigs Taggart: is the goal "resolving bogus issues to keep the noise down"
[15:28] Gigs Taggart: or "finding worthy issues"
[15:28] Saijanai Kuhn: I['ve been workign forever to get my jira suggestion votes stuffed. It aint as easy as you think
[15:29] Rob Linden: "finding worthy issues"
[15:29] Saijanai Kuhn: unless people are creating sock-puppets and voting
[15:30] Saijanai Kuhn: most peole won't vote, regardless. They may complain about a bug, but they don't vote to fix it
[15:30] Rob Linden: the "finding worthy issues" is a process that's really valuable to have multiple Lindens involved in, especially others have already engaged in the "reduce noise" exercise
[15:30] Dale Glass: I vote, sometimes
[15:30] Gigs Taggart: Well, he's tabled patches too.
[15:30] Gigs Taggart: I think patch should automatically be worth a linden look.
[15:30] Rob Linden: I'll import a bug that I see as having a clear repro and very plausible
[15:31] Rob Linden: Gigs: agreed on patches
[15:31] Rob Linden: if there are any patches that have been tabled....feel free to untable
[15:31] Gigs Taggart: ok
[15:31] Gigs Taggart: we are verging on edit war mode here though Rob :)
[15:32] Stevex Janus: I think that we (as programmers) are lazy and don't look at other people's issues. Let alone vote.
[15:32] Gigs Taggart: I untabled everything and he reverted that edit
[15:32] Gigs Taggart: I don't understand his intent anyway, there's nothing on the agenda that isn't tabled right now
[15:32] Rob Linden: did everything have patches attached?
[15:32] Gigs Taggart: no, but the entire agenda is tabled, so I don't understand
[15:32] Rob Linden looks
[15:33] Gigs Taggart: just going to sit around monday and shoot the breeze? :)
[15:33] Rob Linden: no....we'll go through the "tabled" that may be mislabeled
[15:33] Stevex Janus: Or we can just vote.
[15:33] Rob Linden: it may make sense to treat the two sections as different priority levels
[15:34] Rob Linden: Stevex, that works too
[15:34] Rob Linden: here's what I'm hoping for
[15:34] Stevex Janus: Is there a way to speed up the machine running jira? Its slow too.
[15:34] Gigs Taggart: Dzonates has put in a lot of effort and I think he's mostly doing a good job, but I keep running into his stubborn streak :)
[15:34] Rob Linden: I know there's more that we should be importing. at first, I was importing a lot, and I'd get a lot of issues closed out
[15:35] Saijanai Kuhn: may be a security issue. The rest of the forum site is reasonably fast
[15:35] Rob Linden: so, we started doing triages....and I was finding that me setting the order wasn't working so well
[15:35] Rob Linden: I've generally felt the community constructed agendas have been much more productive to walk through
[15:36] Gigs Taggart: yeah I just think we need to probably come up with some guidelines now so there is less confusion like this
[15:36] Rob Linden: sure....suggestions?
[15:36] Gigs Taggart: well we can start with patches being always considered top priority
[15:36] Gigs Taggart: that seems pretty uncontroversial
[15:36] Rob Linden: I'd like to be respectful of everyone's time, not just Linden's
[15:36] Rob Linden: yup
[15:36] Stevex Janus: We could setup teams to lok at jira
[15:37] Gigs Taggart: I think considering the reliability of people on SL, including myself, a loose collaboration is best Stevex :)
[15:37] Stevex Janus: :-)
[15:37] Gigs Taggart: but some guidelines wouldn't hurt I'd say
[15:37] Rob Linden: sure
[15:37] Gigs Taggart: maybe you could put these on the main triage page rob
[15:38] Stevex Janus: Rob, thanks for creating the Source code catagory.
[15:38] Gigs Taggart: Another one could be unresolved unimported bugs with over 10 votes get looked at
[15:38] Saijanai Kuhn: 10 votes is a LOT
[15:38] Rob Linden: hmmm.....I'm happy to referee the construction of guidelines, but I don't want to be the one to construct the guidelines, because we've already determined I suck at putting the list together
[15:39] Gigs Taggart: it is, but big issues get 10 votes quickly
[15:39] Gigs Taggart: this is just the "high priority" category
[15:39] Saijanai Kuhn: ah, right
[15:39] Rob Linden: anyone object to my posting these minutes?
[15:39] Gigs Taggart: go ahead
[15:39] Stevex Janus: nope
[15:39] Dale Glass: no
[15:40] Rob Linden: k...let's have this conversation, I'll post the minutes, and then I'll look for a volunteer to acutally construct guidelines consistant with these minutes
[15:40] Gigs Taggart: k
[15:41] Rob Linden: if an issue has >10 votes, and hasn't been imported, we should talk about it, no question
[15:41] Rob Linden: (assuming, of course, that my belief abou tthere being <10 issues week that clear that bar stands true)
[15:41] Gigs Taggart: I think we should give attention to bugs with solid repros, regardless of votes... something like a rare crash can fall into that, but if the repro is 100% then it's pretty easy to fix.
[15:42] Gigs Taggart: yeah there's usually only 2 or 3 high voted unimported issues
[15:43] Rob Linden: I also think your "solid repro" criteria is a good one, with the caveat that I'm wondering how that gets determined in an edit-war proof way
[15:43] Gigs Taggart: usually the repro itself isn't controversial, but the severity might be
[15:43] Gigs Taggart: by saying "solid repro=triage" then that takes the controversy over severity off :)
[15:44] Rob Linden: I'm willing to give that a shot
[15:44] Rob Linden:'s how I'd really like to see thigns go
[15:45] Rob Linden: there are a number of people I generally trust to just tell me "yes, this is obvious". all of the regulars that are here right now qualify
[15:46] Stevex Janus: Does it count if you are sleeping? : -)
[15:46] Rob Linden: if there's a bunch of issues I can blast through without really having to have a triage discussion, I'd be happy to do that
[15:46] Gigs Taggart: well you could take that to an extreme, and FICify some people to be able to set a special triage Jira flag
[15:46] Gigs Taggart: but that may cause backlash :)
[15:47] Rob Linden: I wish I knew how to set permissions on a per field basis
[15:47] Gigs Taggart: hmmm
[15:47] Gigs Taggart: it probably can be done I've been told Jira has insane permissions detail :)
[15:47] Rob Linden: I've been told that too, but I've done a cursory look for this, and can't find it
[15:48] Rob Linden: I haven't done enough investigation to definitively say it can't be done
[15:48] Gigs Taggart: yeah
[15:48] Gigs Taggart: even if it was a public flag, edit-wars aside, it would be easier than the wiki agenda probably
[15:48] Rob Linden: anyway, what Dz did on the wiki in the lines below the issues is insanely helpful
[15:49] Gigs Taggart: yeah he's doing some good work
[15:49] Gigs Taggart: he resolved about 10 of the issues too
[15:49] Rob Linden: to the extent that we can do an asynchronous "triage" on the wiki, we should try to do it
[15:50] Saijanai Kuhn: on that subject, I've got a regor suggestion for jira itself
[15:50] Rob Linden: that the real-time version is much more about going through issues that really need a discussion to make a determination
[15:50] Rob Linden: Saijanai....go ahead
[15:51] Saijanai Kuhn: there's a huge number of scripting issues scattered in three categories: VWR, SVC and MISC. About 150-200. It's as many as any of the other major categories if you eliminate scripting from them
[15:51] Gigs Taggart: ohhh yeah, I agree with LSL category :)
[15:51] Saijanai Kuhn: I'd like to see SCRIPT be a top-level category, at least as far as *reporting* bug-fix/suggestions go
[15:51] Fremont Cunningham: LSL - me 3
[15:52] Saijanai Kuhn: there could be a VWR sub-category for things that the viewer team sholud look at
[15:52] Saijanai Kuhn: the rest should be organized along the lines as the bsic categories as the LSLwiki
[15:52] Rob Linden: Hmmm
[15:53] Rob Linden: that seems like it's clearly SVC to me
[15:53] Saijanai Kuhn: I found about 5-10 duplicates in each of the cats
[15:53] Gigs Taggart: it's not Rob, the compiler is client side
[15:53] Saijanai Kuhn: its not about which category it belongs in from theLinden side. Its about how they user reports bugs.
[15:54] Saijanai Kuhn: and suggestions
[15:54] Fremont Cunningham: Indeed!
[15:54] Gigs Taggart: yeah it would make it easier to find dups too
[15:54] Fremont Cunningham: As a user with a LSL bug, I look at the current Jira system.. and give up.
[15:54] Saijanai Kuhn: yep. I put it in MISC than discovered SVC then discovered almost as many LSL in MISC anyway
[15:55] Rob Linden: Saijanai: I'll think about it. I agree that our current taxonomy isn't perfect, but I'm really hesitant to make big changes to it without really giving it a think-through
[15:55] Rob Linden: what I'd like to do is this...
[15:55] Saijanai Kuhn: I understand. But its a real headache
[15:56] Wyn Galbraith shows signs of life.
[15:56] Rob Linden: could you file a "feature request" under WEB/, with this request? Ideally, it'd be a link to a detailed writeup on the wiki
[15:57] Rob Linden: ...but even a basic outline of the idea on JIRA is fine
[15:57] Saijanai Kuhn:
[15:57] Saijanai Kuhn is faster than light
[15:57] Rob Linden: heh, cool :)
[15:58] Saijanai Kuhn: And I haven't been evangelizing this one, so those 4 votes are pretty much "real"
[15:58] Rob Linden: I'm also looking for what the component names of the new LSL project would be
[15:58] Rob Linden: (I believe it's "project" rather than category in JIRA nomenclature)
[15:59] Stevex Janus: Is there a notify option in jira?
[15:59] Saijanai Kuhn: that's what holds peole up. As a project, it belongs mostly in SVC, but as a report category?
[16:00] Saijanai Kuhn: if you took it out of svc, SVC would be roughly 1/2 smaller
[16:00] Saijanai Kuhn: VWR would have 30 less
[16:00] Saijanai Kuhn: I can't even think about MISC
[16:00] Celierra Darling waves
[16:00] Rob Linden: Stevex: yes, there is, but it's not enabled right now....long story, will be fixed soon with any luck, but can't make promises
[16:00] Stevex Janus: ok
[16:01] Stevex Janus: If you are going to change things around then you might want to tell people about their issues mving around :-)
[16:01] Rob Linden: yeah, I know
[16:01] Saijanai Kuhn: that's the big drawback
[16:01] Gigs Taggart: LSL components could be math, functions, syntax, prim properties, compiler, interpreter
[16:01] Rob Linden: well, the good enws is that redirects are automatic
[16:02] Fremont Cunningham adds support anc comment to '153
[16:02] Saijanai Kuhn: thanks
[16:02] Saijanai Kuhn: I'll be putting up a suggestion box sign in the sandboxes for my various jira thigns, BTW.
[16:02] Rob Linden: at any rate,t here's a lot of work designing a new hierarchy, and it's work that I personally don't have the time to do. if there's a proposal suitable for execution by a trained monkey, then it will be more likely to happen
[16:03] Saijanai Kuhn: It catches people's attention becasue it is unique, afaik
[16:03] Rob Linden: since, while I'm not as well-trained as a trained monkey, I'm almost as good at improvising
[16:03] Saijanai Kuhn: could an infinite number of Rob LIndens do the same as the monkeies?
[16:04] Ash Qin: Heh
[16:04] Rob Linden: it'd be close
[16:04] Stevex Janus: The main limitation is the single keyboard they would have to use.
[16:04] Saijanai Kuhn: extreme programing writ large...
[16:04] Celierra Darling: They'd have to construct an infinite number of randomly forest-dwelling cubicles first.
[16:05] Rob Linden: ok....I should wrap up now. anything quick before I go?
[16:05] Stevex Janus: nope.
[16:05] Stevex Janus: thanks for having the meeting.
[16:05] Wyn Galbraith shivers, forest dwelling cubicles.
[16:05] Saijanai Kuhn: noe and thanks for listening.BTW, would voice work i this meeting?
[16:05] Ash Qin: Indeed, spasiba :)
[16:05] Saijanai Kuhn: practically speaking
[16:05] Wyn Galbraith: Can't log voice. ;)
[16:05] Gigs Taggart: can't use voice in linux
[16:06] Rob Linden: I don't think this region is voice-enabled yet
[16:06] Saijanai Kuhn: was just wondering if this was a voice-useful meeting, not whether or not voice works here yet
[16:06] Wyn Galbraith: It will make it hard to post a log, think about it, you would have to repeat every question and answer in chat.
[16:06] Ash Qin doesn't think voice is good for meetings such as these.
[16:06] Wyn Galbraith: Kinda spoils the usage of voice.
[16:06] Gigs Taggart: I'm not rebooting and installing SL in windows just for this meeting :)
[16:07] Stevex Janus: me neither.
[16:07] Ash Qin: The moment you use voice, everyone has to use voice
[16:07] Saijanai Kuhn: taking notes
[16:07] Wyn Galbraith: No one wants to hear me.
[16:07] Fremont Cunningham: Voice meetings need a minutes taker, proposals, voting and so-forth. Real records just like RL
[16:07] Rob Linden: we'd need a notetaker. I'd acutally like to try the occasional voice meeting, but i'd also rather wait until we have Linux support at least
[16:08] Stevex Janus: Is there a 'record' option? for voice?
[16:08] Dale Glass: Rob, is there source for the voice stuff?
[16:08] Ash Qin: Not that I can see
[16:08] Rob Linden: we have voice meetings all of the time for Linden business, and it works remarkably well
[16:08] Fremont Cunningham: Run GoldWave
[16:08] Dale Glass: It probably should just be recorded and uploaded somewhere. Then somebody can transcribe it
[16:08] Rob Linden: Dale, no, not yet. we're still getting our act together on that front
[16:08] Saijanai Kuhn: but lindens know each other, mostly, right?
[16:08] Celierra Darling: Would anyone actually volunteer to transcribe everything?
[16:09] Rob Linden: I'd want realtime notes.
[16:09] Saijanai Kuhn: would be hard to track who said what, otherwise
[16:09] Stevex Janus: It would have to be timestapmed .
[16:09] Ash Qin: Here let me try..
[16:09] Fremont Cunningham: You need a secretary. Been there, done that. Not again :)
[16:09] Celierra Darling wishes voice was a STT->TTS systm :/
[16:09] Dale Glass: isn't there metadata regarding who is the voice coming from?
[16:09] Rob Linden: one thing I've seen work really well in the past was w3c meetings, where we'd be meeting, but a realtime voice minutes were taken via irc
[16:09] Ash Qin: Here goes... Double the killer delete select all
[16:10] Dale Glass: client could be made to create a good voice log. Something like a file per speaker, plus a log with pointers to the time when it was said
[16:10] Rob Linden: Dale: yes, that's a huge advantage of using SL for a telecon-type use...all sorts of great visual indicators of who is talking
[16:10] Saijanai Kuhn: one thing I wnat for voice is auto-mute for people with voice-lock on
[16:11] Fremont Cunningham: Not mant leave voice on. Right click mute
[16:11] Ash Qin: Not many have voice
[16:11] Rob Linden: I have voice lock on, but I have a mute button on my usb headset that mutes me
[16:11] Joshua Linden: Good news everyone! We're going to putting out a first drop of the shortly.
[16:11] Saijanai Kuhn: griefers like to breath heavy, play loud music, etc
[16:11] Ash Qin: I think the common man is a lot more annoying
[16:11] Fremont Cunningham: what hath '18?
[16:12] Ash Qin: The man who has a bad soundcard and bad microphone, constant buzzing, the man who uses speakers instead of headset, so you constantly hear reverb etc.
[16:12] Stevex Janus: ok, got to go now. SSe you on monday.
[16:12] Saijanai Kuhn: laters
[16:12] Ash Qin: See you, Stevex
[16:12] Dale Glass: I'd be guilty of that if I used voice. Although I use headphones
[16:13] Rob Linden imagines a Futurama reference in Joshua's comment, even if one wasn't intended
[16:13] Ash Qin: I actually prefer talking to the griefers on yahoo voice chat and PoW, because they have a clear microphone.. they just have their voices distorted to sound like a computer or something
[16:13] Joshua Linden: Big thing in 1.18 is the work Zero Linden has mentioned about moving from UDP-based messaging to TCP -based messaging for some initial set of messages.
[16:13] Saijanai Kuhn: some people have legit uses for voice lock (besides you, Rob) like lecturers, tour guides, maybe muricians...
[16:13] Joshua Linden: Strictly speaking, LLSD over HTTP
[16:13] Gigs Taggart: cool
[16:13] Rob Linden: ah...righto
[16:14] Ash Qin: Which messaging is moving over to TCP exactly?
[16:14] Fremont Cunningham: More security for the transport?
[16:14] Saijanai Kuhn: prays that group IM gets it fast
[16:14] Rob Linden: more security, and a lot more protocol stability
[16:14] Ash Qin: Well, if by security you mean reliability
[16:14] Rob Linden: (stability in the backwards compatibility perspective)
[16:14] Ash Qin: Although I guess security could be done if it were over HTTPS
[16:14] Joshua Linden: Yeah. The implementation allows individual messages to be flipped on-the-fly (on the server). So we may experiment a bit to see which make sense to move to TCP.
[16:14] Fremont Cunningham: I do mean reliability :)
[16:15] Joshua Linden: I think we suspect "everything but object updates" will eventually be TCP but we'll see.
[16:15] Saijanai Kuhn: a suggestion for the beta viewer: give everyone a game-wide group so that you can stress test group IM automatically
[16:15] Fremont Cunningham: object position update? Thats the one everyone bitches about :)
[16:16] Celierra Darling: Oh, if there's a quick answer - I saw a mention of a "3 repro, 3 votes" rule...what would count as 'a repro'?
[16:16] Joshua Linden: ... and because new messages and changed messages will be LLSD vs. some funky binary packing, it means backwards/forwards compatibility of viewers vs. the grid. At least as easily as, say, V1 of an app can open documents created with V2 and vice versa.
[16:17] Wyn Galbraith: Time to go.
[16:17] Saijanai Kuhn wants mutiple descriptors per prim and a changed flag for same so we can do flags between scripts
[16:17] Dale Glass: that sounds very good :-)
[16:17] Gigs Taggart: Celirra I don't think we've ever used a 3 repro rule :)
[16:17] Gigs Taggart: 1 good repro is enough, after all
[16:19] Rob Linden: thing that's tough about 1 repro is that the one repro could be some funky config problem for that person
[16:19] Gigs Taggart: I mean a confirmed repro
[16:20] Ash Qin: It would be nice if SL had a method for clients that cannot use UDP for whatever reasons to fall back on something that can get past certain annoying setups
[16:20] Gigs Taggart: that someone has actually used to repro it :)
[16:20] Rob Linden: independently confirmed repro
[16:20] Gigs Taggart: yes
[16:20] Rob Linden: yup, that's good enough
[16:20] Celierra Darling: So repro = confirmation? Or repro = a series of steps? (And of course, what about stuff like "this thing on the website is spelled wrong"? :P )
[16:20] Fremont Cunningham: If you cant repro a bug, write back to the reporter and tell em so, ask for more info.
[16:21] Gigs Taggart: the series of steps is the repro, confirmation is when someone else gets it to happen following those steps
[16:21] Rob Linden: Celierra: obvious stuff like that is fair to import without a "repro"
[16:21] Gigs Taggart: at least, that's the termination I use
[16:21] Gigs Taggart: terminology rather
[16:21] Rob Linden: (though really, a spelling error is easily "reproed"...."yup, spelled wrong in my browser too!"
[16:22] Celierra Darling nods, "That's why I was confused about '3 repro'... I think of 'repro' as the series of steps, too."
[16:22] linux soda machine Mark II v1.05: Here's your can of Tiny flavored Linux soda, enjoy!
[16:23] linux soda machine Mark II v1.05: Here's your can of Tiny flavored Linux soda, enjoy!
[16:23] linux soda machine Mark II v1.05: Machine is in use by Celierra Darling
[16:23] linux soda machine Mark II v1.05: Machine is in use by Celierra Darling
[16:23] Celierra Darling: Doh >.<
[16:23] Gigs Taggart: heh a soda machine with session management
[16:23] Gigs Taggart: go LSL!
[16:23] Rob Linden: ok....I really should go now
[16:23] Gigs Taggart: yeah
[16:23] Celierra Darling: Of course, I clicked 'ignore' instead of 'cancel' :P
[16:23] Gigs Taggart: me too
[16:23] Ash Qin: Have fun, Rob
[16:24] Celierra Darling waves, "Have a nice day Gigs!"
[16:24] Gigs Taggart: later
[16:24] Celierra Darling: And Rob :)
[16:24] Rob Linden: bye all, thanks for coming. have a great weekend!
[16:24] Saijanai Kuhn: take care all