Simulator User Group/Transcripts/2011.03.18

From Second Life Wiki
< Simulator User Group
Revision as of 08:05, 21 March 2011 by Andrew Linden (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Simulator_User_Group

Prev 2011.03.15 Next 2011.03.22

List of Speakers

Andrew Linden Ashiri Sands Cummere Mayo
DogWomble Dollinger Draconis Neurocam Fancy Greeter
Flip Idlemind Hope Dreier Jessica Lyon
Joshua Linden Latif Khalifa Liisa Runo
Peewee Musytari Qie Niangao Rex Cronon
Siana Gearz Sigma Avro SilverbackMale Looming
Simon Linden Stickman Stickman Ingmann
TankMaster Finesmith TheBlack Box

Transcript

[16:02] Draconis Neurocam: hello andrew

[16:02] DogWomble Dollinger waves toandrew

[16:02] Jessica Lyon: Hey Peewee

[16:02] Andrew Linden: Hi everyone

[16:02] Ashiri Sands: hi Andrew

[16:02] Jessica Lyon: Hai Andrew and folks

[16:02] TheBlack Box: hi

[16:02] Peewee Musytari: hey hey andrew

[16:03] Andrew Linden: I think Simon is skiing on some mountain today, so he won't be showing up.

[16:03] DogWomble Dollinger: next time you see him, can you please tell him we're jealous and can we have a group invite next time? :P

[16:03] Andrew Linden: Yeah, I haven't had any snowplay this year either.

[16:04] DogWomble Dollinger: i haven't seen snow in a good 5 years or more

[16:04] Andrew Linden: But it was raining buckets here in SF -- maybe Tahoe is getting lots of good snow.

[16:04] Ashiri Sands tries to remember what snow is like

[16:04] Sigma Avro forgot about sun

[16:04] Andrew Linden: Let me start with the news...

[16:04] Hope Dreier: Ashri, it's white fluggy and very cold

[16:05] Hope Dreier: fluffy*

[16:05] Peewee Musytari: hehe @ fluggy clouds

[16:05] Andrew Linden: We had an emergency deploy Wed night for some permissions exploit.

[16:05] DogWomble Dollinger: just don'r eat it if it's yellow

[16:05] sahara1988: why cant i see anyone :-/ and what is this music about? am new to this...

[16:05] DogWomble Dollinger: what was the perms exploit?

[16:05] Andrew Linden: and threaded-rez is disabled for a different copy exploit

[16:05] Hope Dreier: O-o

[16:05] Jessica Lyon: copy exploit with threaded rez? sigh...

[16:06] Sigma Avro: oh

[16:06] Hope Dreier: it alos makds things hav a date of 0

[16:06] Andrew Linden: which is why SVC-6544 appears to be fixed (was caused by threaded-rez)

[16:06] JIRA-helper: http://jira.secondlife.com/browse/SVC-6544

[#SVC-6544] In "About Land" dialog, "Most Recent" objects show a date of 1970/01/01 00:00 UTC

[16:06] Hope Dreier: nods nods

[16:06] DogWomble Dollinger: aaah interesting find andrew

[16:06] Andrew Linden: Both exploits are fixed, but the threaded-rez one won't be deployed until next week's RC

[16:07] Andrew Linden: in the meantime threaded-rez is disabled

[16:07] Cummere Mayo: ><

[16:07] Andrew Linden: for those who don't know, threaded-rez is an optimized code path for rezzing objects - the asset parsing is done in another thread

[16:07] Andrew Linden: so to have it disabled means a slight degredation of performance

[16:07] Sigma Avro: and first thread is for what then?

[16:08] Andrew Linden: the simulator has a "main thread" that does most of the work

[16:08] Sigma Avro: oh, and second if for textures and so on ?

[16:08] Andrew Linden: SVC-6544 is actually fixed properly in the next maint-server RC (should be next week)

[16:09] Hope Dreier: presumanby the Perms thing too?

[16:09] Andrew Linden: the simulator actually has several threads... four or five, but most of them don't do much... just handle signals from other processes

[16:09] Andrew Linden: the main "worker thread" does the threaded-rez and some threaded work when teleporting or crossing region boundaries

[16:10] Sigma Avro: oh, ok, so thread are more as streams ? sory for my ignorance

[16:10] Andrew Linden: I'm going to start helping the mesh team part-time, starting next week

[16:10] Draconis Neurocam: should be fun for you, no?

[16:10] Andrew Linden: first to make it support prim-encroachment and other things that they haven't had time to make work

[16:11] Fancy Greeter: Joshua Linden has arrived!

[16:11] Latif Khalifa: I didn't realize server side of mesh project was that big

[16:11] Andrew Linden: I will try to get "cross region prim encroachment" done as part of that effort

[16:11] Hope Dreier: woot

[16:11] Peewee Musytari: any idea yet when mesh will be out of beta?

[16:11] DogWomble Dollinger: heyjoshua

[16:11] Andrew Linden: and Falcon and I were wondering today if we could get cross-region collisions working before mesh is delivered. We'll see.

[16:11] Joshua Linden waves

[16:11] Draconis Neurocam: hello joshua

[16:11] Peewee Musytari: heya joshua

[16:12] Andrew Linden: No idea when mesh will be done.

[16:12] Cummere Mayo: hi joshua

[16:12] Jessica Lyon: Andrew that'd be awesome

[16:12] Andrew Linden: My primary goal will be to help get the phys engine upgrade up for RC (with mesh disabled)

[16:12] Peewee Musytari: are u an irish shrimp there josh?

[16:12] Latif Khalifa: Cross region colissions would eliminate the fancy tricks one needs to do when building bridges across

[16:12] Draconis Neurocam: indeed

[16:12] Hope Dreier: A green Nautilus how interesting.

[16:12] Andrew Linden: this will help us deliver all the work that has been done there in smaller chunks

[16:12] Joshua Linden: Nautilus!

[16:13] Peewee Musytari: lol

[16:13] Andrew Linden: Hrm... I think that is all the news that I've got.

[16:13] Liisa Runo: from how far the collision and encoachment works? for example with prims that have the center 6 regions away?

[16:13] Andrew Linden: the table is open for general discussion.

[16:13] Latif Khalifa: that would be nice seems mesh is blocking other projects

[16:13] Latif Khalifa: like the mono upgrade

[16:14] Sigma Avro: yes , encrachemnt on more than 2 regions,

[16:14] Andrew Linden: Liisa, I'd be happy to first get prim-encroachment working just between two adjacent neighbors first.

[16:14] Andrew Linden: Once I get that I'll consider what to do about very distant regions.

[16:14] Liisa Runo: sure, not pushing you, just courious

[16:14] Andrew Linden: However, my guess is that it won

[16:14] Andrew Linden: won't work for distant regions, but I'll probably get it to work across corners.

[16:15] Sigma Avro: corner = 4 regions

[16:15] DogWomble Dollinger: hmmm, i'd like to find out a bit about the mono upgrade project, mainly being once it is deployed, what effect are we likely to see on script performance?

[16:15] Hope Dreier: 3 at most

[16:15] Andrew Linden: yeah, so working for encroaching objects from the NE or SW regions.

[16:15] Andrew Linden: that is my hope, at least.

[16:15] Hope Dreier: nods

[16:16] Rex Cronon: greeting all. my bad for being late

[16:16] Peewee Musytari: heya

[16:16] Latif Khalifa: hey rex

[16:16] DogWomble Dollinger: heyrex

[16:16] Stickman Ingmann: I forgive you, Rex.

[16:16] Rex Cronon: :)

[16:16] Sigma Avro: hi king

[16:16] Andrew Linden: DogWomble, as I understand the mono2-performance project it is mostly to address some lag experienced when rezzing (or region crossing) objects with mono scripts

[16:17] Rex Cronon: this time change is guilty, not me

[16:17] Andrew Linden: the creation of the mono scripts upon arrival of the object takes longer than it should

[16:17] DogWomble Dollinger: aaaah nice :)

[16:17] Andrew Linden: and the mono2-performance project is addressing that

[16:17] DogWomble Dollinger: that sonds like it will help with sim performance when people TP in

[16:17] Andrew Linden: the hope is to make mono scripts perform as well or better than legacy LSL scripts

[16:17] Latif Khalifa: DogWomble, yes

[16:18] Draconis Neurocam: hopefully a step in the right direction for more advanced scripting changes in the future

[16:18] Hope Dreier: And fixing the problm in Mono (1) is non trivially difficult I understand.

[16:18] Andrew Linden: once that is done we'll be trying to identify and eliminate all reaons why anyone would choose to use legacy LSL rather than mono

[16:18] Latif Khalifa: the problem is fixed, just needs to be deployed

[16:18] Andrew Linden: if we could do that then we could start deprecating legacy LSL

[16:18] Joshua Linden: "just" :)

[16:18] Cummere Mayo: wait so old lsl scripts wont run at all?

[16:18] Cummere Mayo: or we just cant make more?

[16:19] Latif Khalifa: Josh, yeah if the mesh people would unblock it ;)

[16:19] Andrew Linden: Latif is correct. the mono2-performance work is done, needs some QA and then deployment into an RC

[16:19] Rex Cronon: should we start sharpening our c#skills?

[16:19] Qie Niangao: (incidentally, Kelly sent an email to a scripting list today: "Good news is a fix for the mono stall bug should be getting out to Aditi by early next week (or even today if I can wrangle it!). Hopefully we can get it to agni in short order and then using mono scripts will be on par with LSL."

[16:19] Liisa Runo: :)

[16:19] Ashiri Sands: good news =^_^=

[16:19] Hope Dreier: Wonderful.....

[16:19] DogWomble Dollinger: very good news

[16:20] Sigma Avro: you will deploy on all SLS ?

[16:20] Sigma Avro is terrified

[16:20] Andrew Linden: Cummere I would expect that we would first make it impossible to make new legacy LSL scripts.\

[16:20] Cummere Mayo: well theres TONS of stuff that would break if you set it so the old ones cant run. I have NO issues with all new ones being mono

[16:20] Andrew Linden: And then we would probably try to figure out how to migrate them to mono

[16:21] Jessica Lyon: migrate.. was hoping you'd say that

[16:21] TheBlack Box: switching off support for existing LSL would be too much of a hassle ... i doubt that will happen

[16:21] Stickman Ingmann: Once mono scripts have no delay and you can set the size of a Mono script, LSL will be less appealing. It still does some things in a smaller footprint, though. So it may still be attractive in certain situations.

[16:21] Latif Khalifa: Stickman, no it doesn't. It's just reported that way

[16:21] Latif Khalifa: common misconception though

[16:21] Peewee Musytari: there are too manyold scripts where the creators have left SL to just break them all isn`t there?

[16:22] Hope Dreier: I expect that for the foresibl futher that the Lsl interpreter will need to be around, until the byte code can be migrated.

[16:22] DogWomble Dollinger: well the only thing that annoys me is if a creator has set the script to no-mod, you can't recompile to mono myself

[16:22] Andrew Linden: Right Stickman. We would want to identify all the reasons why anyone would choose legacy LSL over mono.

[16:22] Qie Niangao: Latif, it does. Some storage in Mono is incredibly bloated.

[16:22] Andrew Linden: And then hammer those reasons down.

[16:22] DogWomble Dollinger: onyl because i've got a few things rezzed out that might see a performance benefit if i recompiled, but i'm not able to

[16:22] Stickman Ingmann: Latif, certain things take up considerably less memory in Mono. The same script in LSL could be considerably smaller than a Mono script that does the same thing, even with proper reporting.

[16:22] Stickman Ingmann: Er. Other way around.

[16:22] Stickman Ingmann: Sorry, just woke up.

[16:23] Joshua Linden: There are also some behavior differences in the runtimes. Some content may be dependent on the quirks of each. Deciding what to do in those cases would be fun.

[16:23] Flip Idlemind: Here's a reason: LSL takes up 16kb, while Mono seems(!) to take up 64kb (even though that's not actually true, we have no way to tell how much memory they actually use)

[16:23] Latif Khalifa: Andrew, the main reason so far was making teleport slow if mono scripts were in attachments. That's what mono2-perfomance is supposed to have solved

[16:23] Andrew Linden: I'm not opposed to adding new LSL functions that only work in MONO, for example, to add incentives for people to move over.

[16:23] DogWomble Dollinger: good pont flip

[16:23] Andrew Linden: However to tell you the truth, we haven't actually kicked that idea around the lab yet.

[16:23] Stickman Ingmann: I would not hate you for doing so, Andrew.

[16:23] Latif Khalifa: Flip, that's the reporting thing

[16:24] TheBlack Box: Latif: also stuff like projectiles

[16:24] DogWomble Dollinger: andrew, is there any plans to have it set so that mono scripts report their true size, rather than defaul to 64k?

[16:24] Liisa Runo: and some automatic recompile to mono might be a problem too. i know some very old LSO scripts have been compiled when the engine was ignoring some stuff that nowdays refuse to compile. the script still run, but dont recompile

[16:25] Liisa Runo: .. not that i would miss them scripts if they would go boom

[16:25] Andrew Linden: DogWomble, I don't know about that memory reporting issue.

[16:25] Latif Khalifa: there are also scripts where source asset has gone missing from the asset server. if there was ever an automatic migration it would have to be done from the bytecode

[16:25] Sigma Avro: automatic recompilation may brake some contents

[16:25] Andrew Linden: I'll have to ask Kelly.

[16:25] Siana Gearz: aren't there scripts around for which source asset is missing? this would mean would have to decompile the bytecode first.

[16:25] DogWomble Dollinger: no problems :)

[16:26] Latif Khalifa: Siana, stop copying my brain waves!

[16:26] Latif Khalifa reaches fr the tinfoil hat

[16:26] Siana Gearz: Latif, stop copying my brain waves!

[16:26] Andrew Linden: Correct Siana. I don't know the feasability of decompiling the bytecode.

[16:26] Jessica Lyon: lol

[16:26] Siana Gearz: Andrew, do we have the bytecode format?

[16:26] Latif Khalifa: perhaps my eco-facelight will halp

[16:26] Andrew Linden: But I'm just speculating on how we could help convince people to stop using LSL in the future.

[16:26] Andrew Linden: er... "legacy LSL"

[16:26] Jessica Lyon: rofl

[16:27] Siana Gearz: perhaps i can write a bytecode to source decompiler...

[16:27] Latif Khalifa: it uses solar and wind to power my facelight, it's all green!

[16:27] Stickman Ingmann: Nice.

[16:27] Liisa Runo: just prevent us from compiling to LSO in the future

[16:27] Peewee Musytari: i needd to buy new retinas anyways

[16:27] Joshua Linden: anyone have a JIRA on the "mono always shows as 64k" issue? I'm sure I was looking at one on a list this morning....

[16:27] Sigma Avro: give 10 lindens by lsl script abandoned

[16:27] Stickman Ingmann: Force everyone to use Mono and there you go. Only issue then is the reporting will make people freak out.

[16:27] Hope Dreier: Oh that is pretty straightforward, just remove to [ ] lsl on the compile option

[16:28] Stickman Ingmann: How difficult WOULD it be to fix the reporting?

[16:28] Andrew Linden: hehe, that is an interesting idea Sigma. Just pay for the transition.

[16:28] Rex Cronon: which language will we be using with mono?

[16:28] Andrew Linden: I dunno how hard it would be to fix the mono size reporting.

[16:28] Hope Dreier: Is that per script, or per instance of said script?

[16:28] Cummere Mayo: isnt someone already working on fixing the script reporting?

[16:28] Hope Dreier starts copying scripts like crazy

[16:29] Andrew Linden: at the moment mono uses the same LSL syntax, but runs on mono bytecode

[16:29] Draconis Neurocam: a script tax, like the ancient prim tax, hehe

[16:29] Liisa Runo: proper size reporting would be cool, and might also make some mono haters like mono more ;)

[16:29] Latif Khalifa: Josh, it's compliated. It's not only about reporting. It was related to the limits project where you need to know in advance how much ram it's reserved for the script. So in a way report works, it's just shows reserved, and not actually used memory.

[16:30] Cummere Mayo: which is problematic for lots of reasons, but espeically sl sporting events where youre trying to figure out who is lagging the sim

[16:30] Sigma Avro: I wander how one could get a reliable llGetFreeMemory funcrtion, that is realy needed

[16:30] Andrew Linden: The plan used to be to start using C# for writing mono scripts.

[16:30] Andrew Linden: that project is on hold at the moment

[16:30] Andrew Linden: however it depends on the mono2-performance stuff anyway

[16:30] Draconis Neurocam: hopefully not forever

[16:30] Sigma Avro: yeah, the one left

[16:31] Hope Dreier: Pointer type, llMalloc( integer ); and llFree( pointer) ; would do wonders.

[16:31] Andrew Linden: so we'll revisit that idea once this current mono project is done

[16:31] Stickman Ingmann: Kelly mentioned on Tuesday that he's waiting for mono2, small scripts, script limits, and big scripts, and quite possibly several other projects, before C# happens.

[16:32] Latif Khalifa: my understanding was the brighton team were not far off from beta when they were laid off

[16:32] Stickman Ingmann: That's what was said, yes.

[16:32] Latif Khalifa: hoepfully doesn't take too much to resume

[16:32] Stickman Ingmann: Kelly seems to have a different order of priorities.

[16:32] Cummere Mayo mutters about the whole brighten office thing

[16:32] Jessica Lyon: sighs

[16:32] Qie Niangao: I don't think it's priorities; just no resources.

[16:33] Draconis Neurocam: yeah, people power probably

[16:33] Joshua Linden: for something to get done you need both the people and for it to be at the top of the list

[16:33] Draconis Neurocam: with more people, there could be more lists for more things to be at the top of

[16:34] Cummere Mayo: im gonna keep my mouth shut for now about the people part...

[16:34] Flip Idlemind: I can't imagine that "reduce lag" isn't at the top of the list

[16:34] Sigma Avro: I propose parallel doing, hehe

[16:34] Cummere Mayo: the trouble with making everything the top of the list though is then in the end... how do you know what to start?

[16:34] Latif Khalifa: hopefully mono2 rollout fixes the rez speed problem, and we can move from there

[16:34] Rex Cronon: u mean that fix group chat is not first place:)

[16:34] Joshua Linden: I should mention: one of the active projects that hasn't been mentioned is some groundwork for performance (lag-busting) efforts

[16:34] Flip Idlemind: Well "reduce lag" is also super broad

[16:34] Draconis Neurocam: rex theres a test viewer out for aditi for group chat

[16:35] Rex Cronon: heard of it

[16:35] Hope Dreier: WOOTTTTTTT death to chat lag

[16:35] Sigma Avro: whish lag,? chat lag ?

[16:35] Draconis Neurocam: yes

[16:35] Cummere Mayo: unfortunately i dont think chat lag dissapears... does seem better

[16:36] DogWomble Dollinger: well "better" is an improvem,ent on what we have now

[16:36] Sigma Avro: no, and dont' see how to face chat lag in international network

[16:36] Jessica Lyon: xmpp ftw

[16:36] Sigma Avro: well, ok

[16:36] Draconis Neurocam: josh, if it hasnt been mentioned, any info you have for us?

[16:37] Joshua Linden: *I* was referring to simulation lag - we're working to get the tools in place to try and reduce that aspect of lag. i.e. making things rez faster, keeping sim frame rate up, optimize the network protocol, etc. We're doing data gathering first so we know what's the most important to tackle.

[16:37] Latif Khalifa: Josh, did CAPs gzip compression get turned on?

[16:38] Cummere Mayo: what have you found so far?

[16:38] Joshua Linden: Latif: if you mean, inventory-over-HTTP - it's being looked at. It turned out to be non-trivial but of course a huge boost.

[16:38] Sigma Avro: and you started with making an improved metrics , right ?

[16:38] Joshua Linden: So far we've found that we're getting some awesome tools together :)

[16:38] Latif Khalifa: Josh, the last comment seemed to say it was trivial. Binary LLSD was non trivial, gzip was simple server setting.

[16:38] Cummere Mayo: funny...

[16:39] Cummere Mayo: so you havent yet actaully done any datacollection yet?

[16:39] Joshua Linden: latif: yeah, then we dug deeper. There's a proxy in the way that needs to be taught about transfer encodings. :P

[16:39] Latif Khalifa: dang

[16:39] Draconis Neurocam: oouch

[16:39] Cummere Mayo headdesks

[16:40] Hope Dreier shakes her head

[16:40] Latif Khalifa: Josh, I think you should enable it for all CAPs, sumarilly. Most of it is text LLSD xml which an be a big win

[16:41] Joshua Linden: cummere: not on this particular push. we have tons of instrumentation and data collection. this push is about getting the tools in place to reproduce real-world behavior in a test environment, so developers can iterate quickly on testing and fixing

[16:42] Latif Khalifa: that's non trivial. good luck with it

[16:42] Joshua Linden: latif: BTW, "cap" is short for "capability" it's not an acronym. And agreed, most LLSD data sent over HTTP is in XML and gzip would be a win. It also all goes through the same set of proxies, so we need to do that foundational work first.

[16:43] Latif Khalifa: well it's basically a web server downloading xml content. not too unlike the web were it was used with success for ages :)

[16:43] Joshua Linden: yep

[16:43] Draconis Neurocam: i would just hope that any optimization done would be to benefit everyone, rather than the most common use case scenarios, but every one possible

[16:44] Hope Dreier: I'd like to, with your permission Andrew, visit SVC-5927, under some specific conditions it's is a major lag causer.

[16:44] JIRA-helper: http://jira.secondlife.com/browse/SVC-5927

[#SVC-5927] Temp on Rezzed objects get queued

[16:44] SilverbackMale Looming: Hey u guys!!

[16:44] SilverbackMale Looming: :)

[16:44] Andrew Linden: Hope, do you mean you want to talk about SVC-5927 now? Sure.

[16:44] Rex Cronon: hey

[16:44] Joshua Linden: draconis: totally agree. I doubt we'll be looking for fixes that only help 10 specific avatars in one particular dance club!

[16:44] Hope Dreier: Hi Silver

[16:45] Hope Dreier: Thank you Andrew

[16:45] SilverbackMale Looming: oh nic outfit :)

[16:45] SilverbackMale Looming: st patrick would be happy :)

[16:45] Siana Gearz: Silverback, you're a cloooud :)

[16:45] Sigma Avro: so, what is on top: encroachment, improved lag, mesh, raucast ?

[16:46] Andrew Linden: Ok so the SVC-5927 problem is hard to "fix".

[16:46] Sigma Avro: or it is in parallel ?

[16:46] Hope Dreier: The problem is that when it kicks in a bunch of Objects get queueed, the all released at once, which make it kick in again.

[16:46] SilverbackMale Looming: am i fixed?

[16:46] TankMaster Finesmith: si

[16:46] Andrew Linden: The main reason it is there is to prevent the simulator from using too many CPU and file I/O resources on that host.

[16:46] SilverbackMale Looming: every one is a cloud on HIP

[16:46] Hope Dreier: Can it just be disabled, and estat manager control?

[16:46] Peewee Musytari: u have a head now silver :)

[16:47] Andrew Linden: We could devote more cycles to the partuclar simulator process

[16:47] SilverbackMale Looming: lolz

[16:47] Andrew Linden: however it can cause a performance problem on the regions that are sharing the same host.

[16:47] Andrew Linden: If we let estate managers control that particular throttle, then they could tweak things to the detriment of others.

[16:47] Draconis Neurocam: how many regions per host again, ive not heard the latest numbers for a while?

[16:48] Hope Dreier: I'm reasonably sure that when it kicks in and starts cycling eveonr on the host goes to heck all at once, in bursts

[16:48] Sigma Avro: mmh, prefer no hardware neighbours

[16:48] Stickman Ingmann: One per core for full region. Four per Homestead/openspace.

[16:48] Andrew Linden: The simulators had this problem in the past, but we weren't throttling this load back then.

[16:48] Stickman Ingmann: Class 5 has 4 cores, Class 7 has eight.

[16:48] SilverbackMale Looming: is there a way to mute this guy that comes on voice but not in the active speakers list?

[16:48] Flip Idlemind: I thought we didn't have "classes" anymore

[16:48] Sigma Avro: the higher the class, the higher the temp onrez problem then ?

[16:49] Peewee Musytari: I have a performance question about colliders...the debug list used to just show physical objects but there seems to be a lot of strange things on there now...like a reg prim in the corner of the floor in a building or other stuff that couldn`t be colliding with anything...is there a reason for that?

[16:49] Andrew Linden: There may be ways to make the simulator run faster for rezzing lots of objects. Here I'm talking about just general smarter software.

[16:50] Sigma Avro: ok

[16:50] Andrew Linden: I'm in favor or working on such projects, but we can't let estate owners take arbitrarily large server resources.

[16:50] Hope Dreier: what hapens is when Physics drops to a point the temp rezzing gwts queued, then when it rises to a higer t point they all go at once. driving physics down even further.

[16:51] Sigma Avro: is past formula for temp on rez prims amount still valid ?

[16:51] Qie Niangao: so... if the weapons changed simply to be slower to rez (for everybody), the problem would go away?

[16:51] Hope Dreier: Well they only get a certian slice , and the objects tend to be limited to a few specific ones.

[16:51] Sigma Avro: or is there a distinction between phys and not phys temp on rez ?

[16:51] Rex Cronon: slower rezing weapons? u would have an outcry:)

[16:52] Hope Dreier: right now is ~ 1 per second per user, max

[16:52] Sigma Avro: that small ??

[16:52] Stickman Ingmann: My gun needs to shoot 100 bullets a second that do 100% damage and home in on the targets. Otherwise, I'm quitting SL forever. And I make hundreds of Lindens a month, this would ruin you!

[16:52] Hope Dreier: yeah they are bows, not machineguns

[16:52] Sigma Avro: lol

[16:52] Jessica Lyon: I lol'd

[16:53] Hope Dreier slaps Stoickman with her region banhammer.

[16:53] Stickman Ingmann: So Andrew, did you get a chance to talk to Richard yet? What's it been now, two weeks?

[16:53] Cummere Mayo: lol

[16:53] Rex Cronon: unless u r a linden that will only tickle him:)

[16:54] Hope Dreier: The problem is these queues objects are now useless since the target isnlt where he was 20 seconds ago

[16:54] Andrew Linden: Sigh... no Stickman. I have forgotten to talk to Richard about avatar animation.

[16:54] Sigma Avro: all my misiles have to be retought. I would appreciate exact figures about temp on rezes

[16:54] Hope Dreier: lol we're not in Sardar or Esseex either.

[16:54] Sigma Avro: where can I find this ?

[16:55] Stickman Ingmann: Is there a way I could remind or guilt trip you into doing it, Andrew?

[16:55] Hope Dreier: I believe that it's discussed in the JIRA but not in great detail.

[16:55] Joshua Linden: what was the animation question again?

[16:55] Andrew Linden: I don't have a good picture in my mind how the temp-on-rez queuing works and how it might be fixable.

[16:55] Sigma Avro: hands annims ? lol

[16:56] Joshua Linden: hands are for the weak!

[16:56] Andrew Linden: I'd have to run some tests and examine the code some more.

[16:56] Sigma Avro: rezally ?

[16:56] Andrew Linden: However, it will always be possible to *try* to rez faster than the system can support.

[16:56] Hope Dreier: Andrew I, and a few thousand other whould be most greatful if you'd look into it.

[16:56] Stickman Ingmann: The animation question was regarding animation improvements. There were several issues.

[16:56] Sigma Avro: great ! will make avats much less ridi@#{[

[16:56] Latif Khalifa: well temp rezz queuing seem to have become problem relatively recently

[16:57] Stickman Ingmann: Primarily, animation accuracy. Second item was more information displayed for animations -- CPU and memory usage (viewer and server), viewing settings post-upload (ease, duration, frames, etc).

[16:57] Sigma Avro: nice !

[16:57] Hope Dreier: I suspect that the Projectal standards could be rewritten, although weapons makers will have a fit, as will the testers.

[16:57] Sigma Avro: excellent

[16:57] Andrew Linden: yes Latif, where "relatively recently" means since the Havok7 upgrade which happened last summer.

[16:58] Latif Khalifa: it's all Falcons fault ;)

[16:58] Qie Niangao: hmm... well, if weapons-makers get to sell all new weapons, that "fit" might be en route to their banks.

[16:58] Jessica Lyon: gosh, was thatlast summer? already? where does time go

[16:58] Joshua Linden needs to scoot - thanks, all!

[16:58] Sigma Avro: (it is always Falcon fault, I wander why )

[16:58] Draconis Neurocam: take care josh

[16:58] Latif Khalifa: Stickman, the bit about displaying info can be done viewer side.

[16:58] Sigma Avro: thanks to you Joshua

[16:58] Jessica Lyon: Thanks for your time guys.

[16:58] Hope Dreier: 1.42 was the release that broke it

[16:59] Cummere Mayo: thank you all

[16:59] Latif Khalifa: Josh, take care

[16:59] Jessica Lyon: I remember 1.42 well.. still have assets I can't rez >.<

[16:59] Stickman Ingmann: Really? That information is readily available with the animations?

[16:59] Cummere Mayo: i need to take off... after i steal siana's skirt

[16:59] Latif Khalifa: Stick, yes

[16:59] Stickman Ingmann: I don't even know what's possible, honestly.

[16:59] Stickman Ingmann: Trying to figure it all out.

[16:59] TankMaster Finesmith: thx for your time, andrew

[16:59] Sigma Avro: where Latif ?

[16:59] Latif Khalifa: Stickman, the viewer knows all that.

[16:59] Stickman Ingmann: I harassed Zwagoth about making a client that told me information so I could build smart instead of blind, but he was busy.

[16:59] Draconis Neurocam: stickman theres so much animation stuff available thats just embedded in the viewer code

[16:59] Ashiri Sands: Thanks Andrew

[16:59] Sigma Avro: ah, viewer knowns ...

[16:59] Jessica Lyon: lol Stick, Z is busy a lot lately

[16:59] Latif Khalifa: Just need to convice tank or jessica here to make a panel to display it :)

[17:00] Rex Cronon: tc all those leaving

[17:00] Jessica Lyon: lol

[17:00] Stickman Ingmann: He's one of the few people I feel comfortable harassing. Makes it difficult.

[17:00] Hope Dreier: Thank you for your time andrwe, and please if you need some people to hammer as sim let me know, I could probablu get 30 people on 15 minutes notice

[17:00] Jessica Lyon: Z is

[17:00] TankMaster Finesmith: i wondnt know where to get the info to display it

[17:00] Andrew Linden: Yeah, I've got to go now.

[17:00] Andrew Linden: Thanks for coming everyone.

[17:00] Qie Niangao: thanks Andrew

[17:00] DogWomble Dollinger: thaks andrew :D

[17:00] TankMaster Finesmith: have a good weekend andrew

[17:00] Sigma Avro: Thnaks Andrew, nice we

[17:00] Stickman Ingmann: Thanks for the time, Andrew. I'll see if I can email you on Monday. Maybe it'll help.

[17:00] Peewee Musytari: thanks andrew :)

[17:00] Draconis Neurocam: thanks andrew

[17:01] Latif Khalifa: Thank you for your time Andrew, have a great weekend!

[17:01] Jessica Lyon is out for the night. nini folks, have a great weekend!

[17:01] Rex Cronon: have fun:)

[17:01] Andrew Linden: Yeah, send me a reminder Stickman

[17:01] DogWomble Dollinger: and i have to run off as well

[17:01] Latif Khalifa: you too Jessica

[17:01] DogWomble Dollinger waves to everyone

[17:01] Peewee Musytari: oops ended up on table lol

[17:01] Flip Idlemind: Bai guize

[17:01] Andrew Linden: be sure to put "Richard" and "animation" in the subject

[17:01] Latif Khalifa: bye DW

[17:01] Sigma Avro: by all

[17:02] Stickman Ingmann: Will do.

[17:02] Latif Khalifa: i've been collecting facelights lately



Simulator_User_Group

Prev 2011.03.15 Next 2011.03.22