Simulator User Group/Transcripts/2012.07.03

From Second Life Wiki
Jump to navigation Jump to search


Prev 2012.06.29 Next 2012.07.06

List of Speakers

Ashiri Andrew Linden ANSI Soderstrom
Baker Linden dantia Gothly Davido Chrome
Draconis Neurocam Jonathan Yap JS Uralia
Kelly Linden Kennylex Luckless MartinRJ Fayray
Melvin Starbrook Motor Loon Nalates Urriah
NeoBokrug Elytis Sahkolihaa Contepomi Shell Corleone
Simon Linden Squirrel Wood Susan Tsuki
TankMaster Finesmith Warthog Jun Widely Linden


[11:59] Ashiri: Hail Andrew!

[11:59] Warthog Jun: ello drew

[11:59] NeoBokrug Elytis: Hi Andrew.

[12:00] Andrew Linden: Hello

[12:00] Meeter: Welcome to the Server User Group

[12:00] Andrew Linden: Simon may join us later. He's AFK at the moment.

[12:00] Melvin Starbrook: hi andrew :)

[12:00] Susan Tsuki: 'scuse me

[12:00] Melvin Starbrook: hihi hi baker :)

[12:01] Baker Linden: hello

[12:01] Kennylex Luckless: Do not chair eat you this time Baker.

[12:01] Ashiri: Hi Baker

[12:01] Baker Linden: lol

[12:01] Baker Linden: Hi!

[12:01] Susan Tsuki: Jonathan, you closed my bug without measurement

[12:01] Warthog Jun: so lindens

[12:01] Melvin Starbrook: woooot bling hihi

[12:02] Melvin Starbrook giggles

[12:02] Warthog Jun: we have a recurrant inventory problem with shell

[12:02] Shell Corleone: omg

[12:02] Warthog Jun: can you help

[12:02] Simon Linden: Hi everyone

[12:02] Kennylex Luckless: Hi Simon

[12:02] Shell Corleone: it not just me

[12:02] MartinRJ Fayray: Hello

[12:02] Shell Corleone: others get it

[12:03] Shell Corleone: new permanebnt folders

[12:03] Warthog Jun: inventory blowing up

[12:03] Baker Linden: I don't really have any news. I'm still working on fixing management of large groups.

[12:03] Simon Linden: OK, hang on folks - let's do this the usual way

[12:03] Shell Corleone: above and below inventory folder

[12:03] Baker Linden: it's why I've been pretty quiet the last few meetings

[12:03] Shell Corleone: touch them and yu crash

[12:03] Simon Linden: First news -- there was no main channel roll this morning, but that will be on Thursday morning

[12:04] Warthog Jun waves the flag

[12:04] Simon Linden: The blog at has the info

[12:04] MartinRJ Fayray: Shell please go to support for problems with the viewer.

[12:04] Shell Corleone: they do nuffin there

[12:04] Warthog Jun: she has

[12:04] MartinRJ Fayray: then file a jira

[12:04] MartinRJ Fayray: at

[12:04] Shell Corleone: they do nuffin

[12:04] Andrew Linden: There won't be any updates of the RC channels this week (normally on Wednesdays) right?

[12:04] Warthog Jun: again and again and again

[12:04] Sahkolihaa Contepomi: This isn't the place for viewer problems.

[12:04] Warthog Jun: no response

[12:05] Shell Corleone: it isnt viewer

[12:05] Simon Linden: We're here to discuss non-viewer problems here, fwiw ... the Lindens who attend work on the simulator and back-end systems

[12:05] Shell Corleone: i can change viewers

[12:05] Melvin Starbrook: there is a jira allready

[12:05] Shell Corleone: it is still there

[12:05] Melvin Starbrook: this i an old bug im afraid

[12:05] Shell Corleone: so it is old

[12:05] Simon Linden: OK - Andrew, Kelly or Baker - any news?

[12:05] Warthog Jun: and the work around is

[12:05] Shell Corleone: i have to live with it

[12:05] Kelly Linden: No news here.

[12:05] Andrew Linden: Well, if the inventory really gets messed up server-side then it might fall into our domain.

[12:06] Andrew Linden: I'm still working on various bugs in the pathfinding project to get it ready for a real RC channel.

[12:06] Simon Linden: OK, so we're open for topics ... if you have something, the first question from us will be "What's the jira number?" and we're looking for nice detailed problem descriptions so we can get a start on figuring it out

[12:06] Warthog Jun: what is the procedure for reporting this and getting it fixed

[12:06] Andrew Linden: We're supposed to have a code freeze by end of day today for update next week.

[12:06] Shell Corleone: i know many that have it

[12:06] Susan Tsuki: I propose measuring the effect of on a subset of users, blind to the experiment

[12:06] JIRA-helper: [#ARVD-43] stipend tax is flat, not progressive

[12:07] Andrew Linden: And I've got two more bugs to track down.

[12:07] Andrew Linden: As soon as I'm done with that I'm going to start looking into the interestlist with Richard to try to figure out how to fix it

[12:07] Andrew Linden: or rewrite it.

[12:08] MartinRJ Fayray: The procedure is: go to, if they can't help they will tell you to report a bug at All additional explanaitons and how-tos can be found at those two URLs.

[12:08] Andrew Linden: We'll start with some exploratory work and benchmarks to figure out what exactly it is doing.

[12:08] Andrew Linden: That's all the news for me.

[12:08] Warthog Jun: this is a 4 year problem without a resolution

[12:08] Melvin Starbrook: but think it not solved.

[12:09] Melvin Starbrook: trying to find the jira about it...

[12:09] Andrew Linden: I've got a backlog of User Group transcripts that need to be published. I'm going to try to get those done today.

[12:09] Simon Linden: Susan - given the current business climate, there's really no chance that we're going to get into SL taxes. You're free to discuss what it might do, of course, but I'm being realistic about the chances for change

[12:09] Davido Chrome: Bug!

[12:09] Davido Chrome: Squash it!

[12:09] Susan Tsuki: Simon, by "the current business climate," do you mean the declining concurrency and revenue?

[12:09] Widely Linden: its a nice bug!

[12:09] dantia Gothly:

[12:09] JIRA-helper: [#SVC-7940] Get profile image in world never fully loads.

[12:10] Simon Linden: Susan - I mean that in a nice vague, all encompassing way

[12:10] Susan Tsuki: Because that is what it is designed to correct.

[12:10] Susan Tsuki: It could be so important that neglecting it would violate a fiduciary duty.

[12:10] Susan Tsuki: There is only one way to find out: measurement.

[12:11] Susan Tsuki: Or, you could continue LL's longstanding tradition of attempting to sweep unconfortable statistics under the rug.

[12:11] Simon Linden: Dantia - I've reproduced that one, but I don't think anyone has been looking into it. It's in our bug backlog

[12:11] Andrew Linden: Hrm... a quick vote to see how much support the issue might have:

[12:12] Motor Loon: SVC-781 going to roll out soon?

[12:12] JIRA-helper:

[#SVC-781] Allow teleports after region goes down AND/OR always notify agents when region is going down so they have time to teleport away

[12:12] Davido Chrome: Which one?

[12:12] Motor Loon: sorry SVC-7813

[12:12] JIRA-helper:

[#SVC-7813] [PUBLIC]Object Return Function of Region/Estate Menu does not work

[12:12] NeoBokrug Elytis: Baker, I am trying to find the JIRA -- but what's the status of the return all objects in an estate? has it been fixed yet? I haven't had any reason to "test" it. :)

[12:12] Andrew Linden: Anyone here who might be in favor of a tax system being re-added to SL please say "Aye" or something equivalent.

[12:12] Motor Loon: hell no

[12:12] Jonathan Yap: Motor, I brought up the poor visibility of rolling restart messages last week; it is on Andrew's list now

[12:13] Andrew Linden is looking for "Aye's" at the moment.

[12:13] NeoBokrug Elytis: Haha me and Motor Loon are sharing the same brain.

[12:13] Davido Chrome: I abstein on account of economics giving me headaches...

[12:13] Melvin Starbrook: oops o crashed..

[12:13] Motor Loon: so... we each have "half a brain" ?

[12:13] Melvin Starbrook: found the jira

[12:13] JIRA-helper: [#SVC-7381] Unable to Delete Folders Placed Outside of the "My Inventory" Folder

[12:13] Simon Linden: Motor - 78713 has been fixed and is in the release pipeline. It will be at least a couple of weeks before it gets to RC

[12:13] NeoBokrug Elytis: With the way I've been typing today -- most certainly.

[12:13] Warthog Jun: thank you melvin

[12:13] Motor Loon: ok

[12:13] Andrew Linden: Susan Tsuki, as expected we're not finding much support for a tax system in this group.

[12:14] Melvin Starbrook: tax?

[12:14] Melvin Starbrook: i dont like tax lol

[12:14] Sahkolihaa Contepomi would love to see estate return working again. >.<

[12:14] Simon Linden: Speaking of RCs, hopefully some of our bug fixes will be out next week or the following one. The backlog is thinning out

[12:14] JS Uralia: "Allow teleports after region goes down" is a good one. But keeping a backing store on a third server would probably involve latency and reliability issues

[12:14] Andrew Linden: I expect this level of support to be characteristic of the SL resident population.

[12:14] Simon Linden: 7813 will be in the following one

[12:14] Motor Loon: any chance of SCR-111 ?

[12:14] JIRA-helper:

[#SCR-111] llDialog buttons don't say anything after teleporting to another sim

[12:14] Andrew Linden: Also, I would guess there is very little to no interest within LL in adding a tax system.

[12:15] Davido Chrome: Allow Teleports after region goes down? Would need to hand you over somehow? How about autoteleport to an adjacent region?

[12:15] Nalates Urriah: Wait wait... can I tax people if you add a tax system?

[12:15] Jonathan Yap: All this talk about tax is only taxing my patience

[12:16] Melvin Starbrook: mr tax man..

[12:16] Andrew Linden: Right, there are some maintenance projects that are in the queue.

[12:16] Simon Linden: SCR-111 is in our backlog too, but not fixed yet

[12:16] Davido Chrome: Jonathan, calla Taxi!

[12:16] Susan Tsuki: Andrew, I am not proposing "a" tax system, I'm proposing to make the existing tax system more attractive to new and infrequent users, while growing merchant revenue by shifing the burden to the highest earners

[12:16] Davido Chrome groans at his own joke.

[12:16] Motor Loon: Susan want you to pay her more money...

[12:16] Jonathan Yap: At least I was not being teleported to a taxidermist

[12:17] Davido Chrome: I am a socialist, But there is really no need for public healthcare and things like that in SL. Drop it.

[12:17] TankMaster Finesmith: ^

[12:17] Motor Loon: I say lower premium costs and get rid of the stipends

[12:18] Melvin Starbrook: think there are hospitals though

[12:18] Motor Loon: or just keep it how it is...

[12:18] Warthog Jun: As a user tired of having to wait weeks or months for inventory fixes from support (if they bother to fix them at all): I would like the following abilities added into the viewer and the website:

1) ability to move folders from the root folder back into the My Inventory folder

2) ability to fix items that are one thing and have been corrupted so that the system thinks they are something else (ie a clothing item that is now read as a shape).

3) ability to automatically try and find items "missing form the database"

4) ability to hit a button to flag an item for automatic triage and manual fixing when the above steps have not worked.

not only would this greatly improve my experience as a user, these would allow a major source of support tickets, misfiled jira's and so forth to no longer tie up as many Linden Lab, Contractor, and Community resources, and would allow for faster diagnosing and fixing of major inventory issues.

[12:18] Susan Tsuki: I volunteer to forgo all stipends, on any alt, now and forever, if you will just take the idea seriously. That is how confident I am. Neglecting this you might as well torpedo your company, but you can't see it coming because in the thick of things you're like the proverbial frog slow boil

[12:18] Kelly Linden: motor: I will look into scr 111 a bit today, but I suspect it may not be easy to fix.

[12:18] Davido Chrome: Waht do they do, treat bake fail?

[12:19] TankMaster Finesmith: a tax system or no is out of scope for this UG

[12:19] Motor Loon: thanks kelly

[12:19] Warthog Jun: this has been determined to be a server side problem

[12:19] Susan Tsuki: Which UG is appropriate for the lindex?

[12:19] Andrew Linden: Some tools for managing/fixing inventory would be nice.

[12:19] Simon Linden: One little tidbit of news went out recently :

[12:20] Motor Loon: shining is a cool ass project

[12:20] Sahkolihaa Contepomi: Wasn't that the runway project?

[12:20] Davido Chrome nods

[12:20] Simon Linden: it's actually a bunch off cool-ass projects

[12:20] TankMaster Finesmith: yeah, cant wait for shining :D

[12:20] Ashiri: Why the name "Shining" anyway?

[12:20] Sahkolihaa Contepomi: Ohhh.

[12:20] TankMaster Finesmith: cus its new and shiny?

[12:20] Andrew Linden: I don't have a good feeling for the number and characteristics of the inventory bugs we have these days -- I haven't worked on inventory in a while.

[12:20] Motor Loon: "oooh shiny!!"

[12:20] Melvin Starbrook: YAY!

[12:20] Sahkolihaa Contepomi: This is that improvement Oz mentioned.

[12:20] Jonathan Yap: Simon, do you know who wrote that 1st sentence?!?

[12:20] Nalates Urriah: Susan, there is not a UG for that subject. In which case you are left with the JIRA and getting people to vote for it.

[12:20] Simon Linden: From the server-side, there are a few interesting things going on with baked textures and the interest list

[12:20] NeoBokrug Elytis: Here's a scripting issue I ran into the other day specifically with llCastRay. I have a system that migrates objects around my estate, then when it reachese a certain destination it throws a ray down to find a "ground" then move there. I recently transfered a whole regions worth of land (a homestead) to a new resident. After they got the land, any object attempting to cast a ray errored out with RCERR_CAST_TIME_EXCEEDED until the region was restarted. There's no JIRAs on this yet because it's sort of a fringe issue. But it seems that transferring a whole regions worth of land borked llCastRay until the region restarted. (Sorry for the wall of text)

[12:21] Andrew Linden: Warthog, I'll forward your proposal to a dev who knows more than I do about inventory.

[12:21] Andrew Linden takes notes.

[12:21] JS Uralia: Susan, if you really want this, you should set up a twitter petition to @rodvik and @mkapor

[12:21] Sahkolihaa Contepomi: Runitai recently changed how baking works in the viewer, but he's had to disable the 'new' way on AMD graphics cards.

[12:21] Simon Linden: No Jonathan, I had no idea we were maknig that post until a residnet pointed it out to me

[12:21] Motor Loon:

[12:21] JIRA-helper: [#SVC-7603] llCastRay fails, when sitting on object that has crossed simborder

[12:22] Susan Tsuki: give up on working within the system, trying to help a Linden get credit for rescuing their company and go straight to the top, JS?

[12:22] Motor Loon: Neo: apparently it will be resolved when pathfinding goes full grid

[12:22] Kelly Linden: Neo: PATHBUG-77

[12:22] NeoBokrug Elytis: Okay, but it's a lone object without a rider. :)

[12:22] Kelly Linden:

[12:22] Simon Linden: But I'm happy the baked texture pipeline is getting an overhaul - that should really help everyone. The old system solved some big problems, but is just to fragile and I'm very sure we re-bake far more often than needed

[12:22] Motor Loon: probably same bug

[12:22] JS Uralia: If you think it will save them, why do you care whether they or you get credit for it?

[12:22] Kelly Linden: Neo: I would expect pathfinding to resolve your issue.

[12:23] Davido Chrome: A question, how reliable are inter object region say messgaes?

[12:23] Davido Chrome: Messages*

[12:23] Susan Tsuki: I know it will save them, but it just seems more polite to try to get them to reconsider after sweeping it under the rug. If they un-sweep it, then maybe things will be more mellow

[12:23] Simon Linden: with llSay() ?

[12:23] Davido Chrome: llRegionSay

[12:24] Andrew Linden: Right, NeoBokrug, the time allowed for llCastRay calls has been expanded in the pathfinding project.

[12:24] Motor Loon: why wouldnt it be reliable?

[12:24] Simon Linden: Alve - I'm not sure how to measure it, but under normal load, they should be very reliable.

[12:24] Davido Chrome: I made a self setup TP. It works about 50% of the times.

[12:24] Motor Loon: you probably made a boo-boo

[12:24] JS Uralia: Susan, companies aren't turned around by polite people usually

[12:25] Andrew Linden: I would expect llRegionSay() should be very reliable.

[12:25] Andrew Linden looks at the code...

[12:25] Davido Chrome: I might. XD But, yeah. It works HALF of the times!

[12:25] Motor Loon: I've had stuff that would shout over simborders and such from object to object... been running very stabilie.

[12:25] Susan Tsuki: I have no interest in becoming agressive with any of them in any way. If they can't figure it out, then Darwin says they deserve what they get

[12:25] NeoBokrug Elytis: I'll add that this morning I may have had an instance where llRegionSay not work.

[12:26] Simon Linden: There are some internal queues for messages and such, so under high load or bogged down scripts, they probably have a limit and messages will be dropped (I really hope there's a sane cap in there) But otherwise I'd expect it to be 199%

[12:26] Simon Linden: 100% too

[12:26] Motor Loon: lol

[12:27] Davido Chrome: Worked this time.

[12:27] Simon Linden: Do you think you've seeing messages dropped?

[12:27] NeoBokrug Elytis: It always works when you try to show it off. :)

[12:27] Motor Loon: Murphys law

[12:27] Davido Chrome: Hum, this time too.

[12:27] Susan Tsuki: I continue to propose that the effect on user retention and spending of lessening the transaction fee on low income earners and raising it on high income earners should be measured

[12:28] Draconis Neurocam: simon, any word on how exactly it knows if the object cache is valid before it repulls the information?

[12:28] Draconis Neurocam: in the new situation

[12:28] TankMaster Finesmith: it sais in the blog

[12:28] Motor Loon: looks like your 50% successrate aint quite holding up Davido

[12:28] Davido Chrome: No, On three attempts all have succeded. XD

[12:28] NeoBokrug Elytis: Maybe put a thing that counts llRegionSays vs recieved.

[12:28] Draconis Neurocam: tankmaster is says that it checks if it is valid, im wondering what makes it considered valid

[12:29] Simon Linden: Darconis - that's still being sorted out. Those projects are on-going, and we're trying to really nail down the info that the simulator needs to send to teh viewer

[12:29] Andrew Linden: Ah yes, it appears that llRegionSay() events will compete with other llSay() events

[12:29] Draconis Neurocam: right

[12:29] TankMaster Finesmith: the viewer sends a list of objects i knows of with a timestamp, the simulator compairs it to the list of objects it has, then tells the viewer wats still valid and sends the new /updated stuff

[12:29] Andrew Linden: so if there were a flood of chats then the recipient may not recieve them... it has a limited size input queue

[12:29] Draconis Neurocam: simon answered it

[12:29] Davido Chrome: Andrew, what does that mean?

[12:29] Simon Linden: Here's a basic example if you get into the viewer code -- there's a "kill" message sent from the simulator to the viewer. It essentailly tells the viewer "you can't see this object, you don't have to worry about it any more"\

[12:30] Davido Chrome: How biog can the queue be?

[12:30] Meeter: Timecheck : User Group is half over

[12:30] Davido Chrome: Long**

[12:30] Andrew Linden: That means, if your recipient script cannot handle listen() events fast enough for the incoming flood, then some of the incoming messages may get dropped

[12:30] Susan Tsuki: sounds like they use that message a lot in the statistics publication department, Simon >snerk<

[12:30] Andrew Linden: because there is no room in the queue of pending messages to be handled

[12:30] Simon Linden: However, there is no distinction between the object disappearing, and it going out of view. If you can turn around and see it again, we want it cached. If it was deleted from the world, it shouldn't be

[12:31] Davido Chrome: Hum. That shouldn't happen... Each message back is triggered by getting the previous message...

[12:31] Susan Tsuki nods in agreement with Simon

[12:31] Simon Linden: So we're trying to figure out all the extra info the simulator might send, and in the right priority, to let the viewer do the best job

[12:31] Draconis Neurocam: right, i think the issue might come in how much, how often/when and if the object is modified after it has been cached

[12:32] Motor Loon: (you look really cute today btw kelly lol)

[12:32] Susan Tsuki: there are two kinds of statisticians: those who are running towards the data, and those who are running away from it

[12:32] Simon Linden: Some, like that example, seem pretty simple, but even that requires a fundamental change to simulator-viewer messaging

[12:32] Davido Chrome: Ok, I am failing to reproduce... Maybe at the next UG. XD

[12:32] Simon Linden: It's fixed!

[12:32] TankMaster Finesmith: lol

[12:32] NeoBokrug Elytis: hah

[12:33] Motor Loon: You need to add more bugs Davido

[12:33] Ashiri: Alve, were you noticing the bug on a full-speed region or was it idking?

[12:34] Ashiri: *idling

[12:34] Davido Chrome: Ashiri, I am not sure I understand the question, a region idles?

[12:35] Davido Chrome: Ah, there, this time it failed...

[12:35] Andrew Linden: After reading back I see Draconis's comment that opened the discussion about the object cache on the viewer.

[12:35] Motor Loon: so not we just need a solid repro hehe

[12:35] Andrew Linden: Draconis, we're going to be looking into that part of the system over the next few weeks.

[12:36] Andrew Linden: With an eye on how to fix it, or rewrite it.

[12:36] Motor Loon: also part of shiny?

[12:36] Andrew Linden: Not really a part of shiny, but we may be doing some viewer work as part of the project.

[12:37] Motor Loon: cool cool

[12:37] Motor Loon:

[12:37] JIRA-helper: [#SVC-4632] People getting past Estate and Land bans

[12:37] Simon Linden: I'd love to see some better land caching ... it seems the viewer can be pretty slow to draw the ground even in places I've been recently

[12:37] ANSI Soderstrom: Simon Linden is talking about "love"

[12:37] Simon Linden: lol

[12:38] ANSI Soderstrom: sorry O.o

[12:38] MartinRJ Fayray: aren't there plans to replace the current terrain with a mesh?

[12:38] TankMaster Finesmith: i know monty is working on network related protocol/communication changes for between the viewer and the servers

[12:38] Draconis Neurocam: it really makes me happy when the server devs end up having to touch the viewer, because they can see things from both sides

[12:38] Simon Linden: Yes Phoenix - we need to use http and tcp/ip sockets better. That should help with throughput and reliablity

[12:39] Andrew Linden: MartinRJ, the pathfinding project replaces the *collision shape* of the terrain with a mesh (instead of a heightfield)

[12:39] MartinRJ Fayray: ah

[12:39] TankMaster Finesmith: to enabled keep alive and other things that will help to increase communication speed and reliability

[12:39] Andrew Linden: however the streaming of the terrain data to the viewer has not changed yet

[12:39] Davido Chrome: What about sim crossings? They seem to be rather more flaky since the latest update.

[12:39] Andrew Linden: if we do work on that it would be after the "interestlist" work that we're starting now

[12:39] Motor Loon: really?

[12:39] Motor Loon: I've had fine crossings

[12:40] Squirrel Wood: try crossings when you have a script load of 10-20 mebibytes.

[12:40] Simon Linden: There is ongoiing work being done on sim crossings. Initially it's aimed at getting rid of the lag event someone leaving the region causes for everyone else, but is groundwork for faster crossings and TPs

[12:40] Motor Loon: haha thats bound to fail aint it

[12:40] JS Uralia: Susan, if you, IC, were going to be agressive on the issue, what would it be like?

[12:41] Draconis Neurocam: the only issue i have with sim crossings is when i can move into them faster than the terrain loads and have to fly back out and in again

[12:41] Davido Chrome: Oh, there was something I wanted to bring up. would it be possible to implement and option check box for prims that say "do not drop collision events" for prims that aren't supposeded to be roads, but walls?

[12:41] Susan Tsuki snickers

[12:41] Squirrel Wood: Back in ye olde days you could travel the lands and stuff would actually rez ahead of you.. ^^

[12:42] TankMaster Finesmith: they still do

[12:42] Simon Linden: It all depends on how fast you go and where :P

[12:42] Susan Tsuki: Sadly, JS, I'm nicer than Darwin

[12:42] TankMaster Finesmith: and the kind of connection you have with LL

[12:42] TankMaster Finesmith: :D

[12:43] Squirrel Wood: LL needs some european data servers for them european users :p

[12:43] Andrew Linden: What do you mean "drop collision events"?

[12:44] TankMaster Finesmith: with the work monty hopes to accomplish, ISPs will automatically cache textures fetched via HTTP, which will help those on the other side of the world a lot

[12:44] Motor Loon: I haven't really seen noticeable different using Second Life comparing logging in from europe and loggin in from the states... a bit higher ping "on paper" but dont really feel the difference

[12:44] Simon Linden: I believe it was mentioned in that blog post, but we're actually consolidating server centers. The lag between server centers is one of the reasons crossings and TPs are slow

[12:44] TankMaster Finesmith: yes, simon

[12:45] TankMaster Finesmith: from 3 to 2

[12:45] Davido Chrome: Andrew, you guys told me while I was researchign the physics engine that the reason low physics cost physical objects sink into prims are because they drop collision events to make prim roads viable. I want an option to turn that collision dropping off for sopecific prims.

[12:45] Motor Loon: yeah less centers will likely be much better

[12:45] Squirrel Wood: motor, you have about 100ms higher ping times on average and thus requests take longer and thus transfer speeds are significantly lower

[12:45] Motor Loon: but you dont feel it much

[12:45] Squirrel Wood: you do.

[12:45] Motor Loon: I have 180ms ping here from europe... and had like 15ms logging in while in the states... similar performance

[12:46] Motor Loon: Second Life aint that fast anyway ;-)

[12:46] Ashiri: I notice the ping time when driving or flying.

[12:47] Davido Chrome: You told me the reason for dropping collision events was to prevent snagging at seams in prim roads.

[12:47] Squirrel Wood: my connection can do 35mbit. SL is set to use 1.5mbit which is the max you can set it to. It barely reaches 700kbps on average when loading stuff.

[12:47] Motor Loon: yep

[12:48] Motor Loon: Second Life is slow regardless of your connection

[12:48] Andrew Linden: Sorry Davido, was caught uip in some IM's.

[12:48] Davido Chrome: I always have 180 ping. I can still kick american peoples ass in Simball.

[12:48] MartinRJ Fayray: hehe

[12:48] Motor Loon: thats because you cheat Davido

[12:48] Davido Chrome: Pffft

[12:48] Davido Chrome: XD

[12:48] Melvin Starbrook: lol

[12:48] Squirrel Wood: 160ms in team fortress are prefect for sniping. giant hitboxes :p

[12:48] Draconis Neurocam: another thought i had, is you are prebaking outfits, these are going to take into consideration baggy non prim clothing, because there have been issues with them not loading properly in the past

[12:49] Simon Linden: I thought you could get up to about 3mbit, Squirrely - that might be something to check out here

[12:49] Susan Tsuki:

[12:49] Andrew Linden: Davido, we changed some physics engine parameters that should reduce the number of bad normals at road prim boundaries

[12:49] Simon Linden turns the SL server control knob from "medium" to "fast" and sees what it does

[12:49] Motor Loon: I heard it got capped around 1 or 1.5

[12:49] TankMaster Finesmith: viewer max is 10k, simulator max is 3k

[12:49] Andrew Linden: but I don't think they are totally eliminated, especially for concave objects

[12:49] Squirrel Wood: maximum bandwidth can be set to 10000kbps in the viewer I use. It used to be 1500 a year or two ago

[12:49] Simon Linden: Those are the numbers I remember

[12:49] Draconis Neurocam: yeah the viewer max should not be different than the sim max

[12:50] TankMaster Finesmith: but in reality the viewer gets up to only 4.5k internally despite the preference setting

[12:50] Andrew Linden: that is, concave meshy/prim objects for roads and/or moving objects

[12:50] Motor Loon: dont really matter much... slowests will win regardless

[12:50] Simon Linden: However you connect to multiple regions, so you want the viewer to be higher

[12:50] Andrew Linden: but I think the changes we made around collisions at prim boundaries has more to do with the creation ofthe collision events, rather than dropping existing events

[12:51] Davido Chrome: You mean like hollowed squares?

[12:51] Andrew Linden: that is, it happens much earlier, before the collision events are actually created

[12:51] Andrew Linden: but why do you think you need the extra collision events for walls and such?

[12:51] Andrew Linden: are your scripts not getting certain collision events?

[12:51] Davido Chrome: No. Not scripts.

[12:52] Motor Loon: I think he wants no event

[12:52] Motor Loon: s

[12:52] Davido Chrome: The problem is that physical vehicles with low physics cost are prone to go straight through walls.

[12:52] Andrew Linden: ah, well "vehicles" are special

[12:52] Davido Chrome: Unless you go reeeeal slow...

[12:52] Andrew Linden: that is, scripted objects that use the llVehicle* calls

[12:53] Andrew Linden: there is still some special-code hackery around vehicles that WILL drop some real collision events

[12:53] Andrew Linden: and that is what is causing them to go into walls and other obstacles when they move quickly

[12:54] Davido Chrome: Also makes the simboard sink into prim ramps and hit the wrong surface of the next prim in the ramp. And you come to a very abrupt stop.

[12:54] Motor Loon: too small collision models?

[12:54] Andrew Linden: unfortunately, the last time we tried to remove that hackery we found that the vehicles were still hitting the prim boundary normals too hard

[12:54] Motor Loon: havok dont deal well with small stuff

[12:54] Andrew Linden: causing them to jump and go flying at the prim borders

[12:55] Meeter: Timecheck : User Group is almost over

[12:55] Davido Chrome: Andrew, that's why I thought a check box for enabling or disabling said hackery would be nice.

[12:55] Andrew Linden: perhaps it should be revisited, but we'd probably have to get Falcon to revisit the problem

[12:55] Andrew Linden: Hrm.... some way to disable it...

[12:55] Draconis Neurocam: put it in the sim console?

[12:55] Andrew Linden: not a checkbox, but maybe an llVehicle parameter.

[12:55] Andrew Linden: might be possible

[12:56] Motor Loon: heh, I think Falcon has given up on the problem

[12:56] Davido Chrome: Andrew, I would rather have it be a check box tied to specific prims.

[12:56] Andrew Linden: Yeah, Falcon would want to sweep all old style vehicles away and only have Havok vehicles

[12:56] Davido Chrome: So you can have on behaviour on some prims and the other on others.

[12:56] Andrew Linden: Davido, that would be very hard to do.

[12:57] Andrew Linden: That is, per-prim checkbox

[12:57] Andrew Linden: or even a per-prim setting

[12:57] Motor Loon: per object then?

[12:58] MartinRJ Fayray: when we're at it, can we have a checkbox to turn off shadows for a prim?

[12:58] Andrew Linden: It's been a while since I worked on that vehicle hackery code. It is possible that it sn't has hard as I remember.

[12:58] Jonathan Yap: Andrew, will there be a meeting on Friday or is everyone taking a long weekend?

[12:58] MartinRJ Fayray: (so that the prim doesn't cast a shadow)

[12:58] Simon Linden: There should be a meeting on Friday

[12:58] Simon Linden: User group, I mean

[12:58] Andrew Linden: Jonathan, yes there will be a meeting on Friday for anyone with time to attend.

[12:58] Andrew Linden: I'll be working on Friday.

[12:58] Davido Chrome: Andrew, there are other physics settings you can set specifically for different objects. Like friction and bounciness.

[12:59] Warthog Jun: can only be so patriotic

[12:59] Simon Linden: Before we wrap up, I have a little experiment to ask everyone's help with

[12:59] Warthog Jun: perks ears

[12:59] Andrew Linden: Davido, we don't support per-prim friction/restitution do we?

[12:59] Motor Loon: its not something sexual again is it?

[12:59] JS Uralia: Andrew, I've updated the status of the improperly archived bug on and would you please acknowledge that I am proposing to measure changing the existing de facto tax structure, which is flat, to something that gives more incentive for user retention and spending?

[12:59] MartinRJ Fayray: we do, Andrew:)

[12:59] Davido Chrome: Andrew, you support per object.

[12:59] Davido Chrome: And that is enough.

[12:59] Susan Tsuki: Thank you, JS!

[12:59] MartinRJ Fayray: per object, right.

[12:59] Davido Chrome: And that would be enough.

[12:59] Simon Linden: ... I'm going to gather some performance data, and just want people to be active. Move about, maybe do some animations (within acceptable types, of course), rez something, TP away, whatever

[13:00] Andrew Linden: Per-object is possible, but not per-prim.

[13:00] Sahkolihaa Contepomi: :o

[13:00] Meeter: Thank you for coming to the Server User Group

[13:00] Jonathan Yap: say "go!" ?

[13:00] Simon Linden: ok, well, I've started grabbing data - go for it

[13:00] Andrew Linden: JS, I acknoledge that you are proposing a tax structure, but I can assure you there is little to zero support for such a venture.

[13:00] JS Uralia: okay now abduct me so I can live in your socialist utopia, Susan

[13:01] Squirrel Wood: I appear to posess an object that which will not rez

[13:01] Melvin Starbrook: YAY!

[13:01] Kennylex Luckless whispers: ?

[13:01] Sahkolihaa Contepomi: Holy crap

[13:01] Melvin Starbrook: eep

[13:01] Davido Chrome: What happened now? XD

[13:01] Simon Linden: lol no megaprims!

[13:01] dantia Gothly: O_O

[13:01] JS Uralia: Andrew, do you not consider the existing transaction fee a flat tax?

[13:02] Warthog Jun: but its pink

[13:02] Simon Linden: ok, thanks everyone!

[13:02] Kennylex Luckless: What is happening?

[13:03] Davido Chrome: Damn, you have managed to lag this sim down!

[13:03] Simon Linden: That's the idea

[13:03] Squirrel Wood: I wonder if mine scripts here can still slow mono down to a crawl..

[13:03] Simon Linden: But the test is over, so I'm going to start returning stuff

[13:03] Kennylex Luckless: This do scare me.

[13:03] Nearby Stars to 10 parsecs: placing stars...

[13:03] Nearby Stars to 10 parsecs: 10...

[13:03] Motor Loon: most.... interesting...

[13:04] Motor Loon: wait... who farted?!

[13:04] Davido Chrome: Cool. Linden authorized Griefing? XD

[13:04] Melvin Starbrook: ;p;

[13:04] Simon Linden: ok, thanks for playing :)

[13:04] Squirrel Wood: anytime

[13:04] Motor Loon: HIGHSCORE!

[13:04] Warthog Jun: tp failed

[13:04] Squirrel Wood: just holler if you need more help :)

[13:05] TankMaster Finesmith: that was... interesting

[13:05] Simon Linden: That will be interesting performance data -- it actually seemd to hold together pretty well

[13:05] Sahkolihaa Contepomi: Andrew holding a crayon gun...

[13:05] Warthog Jun: failed again

[13:05] TankMaster Finesmith: group member list loaded well on a group with 16k members

[13:05] Motor Loon: so.... Second Life is not unbreakable?

[13:05] Sahkolihaa Contepomi: Reminded of Selkit's massive crayon launcher.

[13:05] Squirrel Wood: fun thing is I have a script that trigegrs a color change of a prim, yet said change is never carried out

[13:05] Squirrel Wood: triggers

[13:05] TankMaster Finesmith: my 118k inventory loaded fine

[13:05] Simon Linden: I'm going to head out - thanks again, have a great holiday (if you have one)

[13:06] MartinRJ Fayray: goodbye

[13:06] TankMaster Finesmith: tc simon, hf!

[13:06] llGetDisplayName: hi all

[13:06] Melvin Starbrook: hihi come to one of our shows lol, we love to make lag too :P

[13:06] ANSI Soderstrom: Melvin Starbrook is talking about "love"

[13:06] JS Uralia: lol

[13:06] Kennylex Luckless: anvil

[13:06] Squirrel Wood: state_entry()










[13:06] Motor Loon: alrighty

[13:06] Melvin's Windup key TO: Melvin Starbrook's windup key has run out.

[13:06] Motor Loon: tc all

[13:06] ANSI Soderstrom: o.O

[13:07] Sahkolihaa Contepomi: Fallout New Vegas Ultimate Edition finished downloading after five hours.

[13:07] Sahkolihaa Contepomi: Time to go play that.

[13:07] Kennylex Luckless: Shall we go bash boxes?

[13:07] ANSI Soderstrom: i am a bit hungry

[13:07] Space Pod v1.1: Say help to a list of commands

[13:07] Kennylex Luckless: Go eat then, eat is good to.

[13:07] Melvin's Windup key TO whispers: Melvin Starbrook winds up the key unassisted.

[13:07] ANSI Soderstrom: lets kill cargo boxes after a biiiiig dinner ;)

[13:07] Melvin's Windup key TO whispers: Melvin Starbrook winds up the key unassisted.


Prev 2012.06.29 Next 2012.07.06