User:Andrew Linden/Office Hours/2010 08 06
Jump to navigation
Jump to search
Transcript
| [16:01] | Ardy Lay: | Hi Andrew |
| [16:01] | Liisa Runo: | /me holds ❤ |
| [16:01] | Techwolf Lupindo: | The last test, I found out http texturts has greater then 1Mbs bandwith. |
| [16:01] | Moon Metty: | hey andrew |
| [16:01] | Techwolf Lupindo: | Hi andrew. |
| [16:01] | Psi Merlin: | Hello Andrew |
| [16:01] | Sindra Aquila: | hello andrew |
| [16:01] | Psi Merlin: | Hi Tech', moon |
| [16:01] | Andrew Linden: | Hello everyone. |
| [16:02] | Moon Metty: | hi simon :) |
| [16:02] | Andrew Linden: | Ok some news/announcements... |
| [16:02] | Techwolf Lupindo: | /me smiles |
| [16:02] | Andrew Linden: | I had mentioned a viewer caching bug that we recently fixed |
| [16:03] | Andrew Linden: | It is currently fixed in SnowGlobe, and will be in the Viewer-2.1.1 release |
| [16:03] | Andrew Linden: | I've asked the viewer team to also fix it in viewer-1.23 because it will also reduce simulator lag a little bit |
| [16:04] | Moon Metty: | nice Andrew |
| [16:04] | Andrew Linden: | however they don't want to do it, the main reason being they don't have time |
| [16:04] | Simon Linden: | hello |
| [16:04] | Moon Metty: | yes, time is short |
| [16:04] | Sindra Aquila: | hi simon |
| [16:04] | Ardy Lay: | Hi Simon |
| [16:04] | Morgaine Dinova: | /me waves the coffee mug at Andrew and Simon |
| [16:04] | Moon Metty: | but is it that hard to fix? |
| [16:05] | Andrew Linden: | while the bug is trivial to fix, their process of releasing stuff is not flexible enough to let them do it on a timely basis |
| [16:05] | Moon Metty: | it's only a typo |
| [16:05] | Kaluura Boa: | Very hard... Must remove a "s" |
| [16:05] | Andrew Linden: | as I said: "trivial" |
| [16:05] | Ardy Lay: | Does Viewer 1.23.5 have any advantages over Snowglobe 1.x? |
| [16:05] | Andrew Linden: | Some people think so, I assume. |
| [16:05] | Morgaine Dinova: | Andrew: if the process for releasing stuff involving a 1-character change is a barrier, the process needs to be changed. |
| [16:06] | Techwolf Lupindo: | Nope. snowglobe has faster UDP texture loading and 1.23.5 does not have http textures. |
| [16:06] | Andrew Linden: | Anyway, now that the answer is final that it won't be included in 1.23 I'll open a jira about it. There is a theoretical workaround for the adventurous who really really want to use 1.23 but reap some of the benefits of the bug fix. |
| [16:06] | Ardy Lay: | Tech, any significan UI differences that might be noticed? |
| [16:06] | Andrew Linden: | I haven't used SnowGlobe. The UI is similar to 1.23? or to 2.X? |
| [16:07] | Techwolf Lupindo: | snowglobe has webkit and media impromvents. |
| [16:07] | Sindra Aquila: | 1.23 |
| [16:07] | Latif Khalifa: | Andrew, snowglobe 1,x has 1.23 ui |
| [16:07] | Andrew Linden: | Oh really? Well then, everyone on 1.23 should be using SnowGlobe, right? |
| [16:07] | Latif Khalifa: | pretty much |
| [16:07] | Ardy Lay: | I think so. :-) |
| [16:07] | Sindra Aquila: | has any been done about the restitution of material yet? |
| [16:08] | Latif Khalifa: | i'd recommend everyone try 1.5.x dev snapshot from https://wiki.secondlife.com/wiki/Snowglobe #Development_Snapshots |
| [16:08] | Latif Khalifa: | it has caching bug fixed + support for http textures with 1.23 interface |
| [16:08] | Moon Metty: | ok Latif |
| [16:08] | Andrew Linden: | That is, is there any reason for a 1.23 user to NOT use SnowGlobe? (I realize this is a rewording of a previous question by someone else) |
| [16:08] | Latif Khalifa: | Andrew, I don't think there is |
| [16:08] | Andrew Linden: | Ok, thanks for the info. |
| [16:08] | Andrew Linden: | Alright... next item... |
| [16:08] | Moon Metty: | well, i was using SG and noticed short client freezes that i didn't see on 1.23 |
| [16:09] | Ardy Lay: | Only thing I can think of is parcel media playback is done by completely different code. |
| [16:09] | Andrew Linden: | A fellow Linden was asking if we could fix SVC-93 today. And also SVC-129. |
| [16:09] | ROTATION and | PRIM_LOCAL_POSITION to complement llGetLocalRot() and llGetLocalPos() |
| [16:09] | ROTATION and | llSetRot incorrectly implemented for child prims |
| [16:09] | Kaine Lowenstark: | considering parcel media isnt exactly working well on 1.23... |
| [16:09] | Morgaine Dinova: | Since 1.23 is no longer supported by LL as it receives no more maintenance, perhaps it should be removed from downloads, and Snowglobe-1.x releases put in its place. |
| [16:10] | Andrew Linden: | I let him know that I've got a fix for SVC-93, and the "solution" is basically to implement SVC-129. |
| [16:10] | Andrew Linden: | I need to do the final review, but I expect to finally get to that next week. maybe it will be in 1.44 (it could happen). |
| [16:10] | Moon Metty: | it's good to see it fixed |
| [16:11] | Techwolf Lupindo: | svc-93 got closed. |
| [16:11] | Andrew Linden: | I'll be adding comments to those two jiras soon (next week). |
| [16:11] | Moon Metty: | rotations are confusing enough without bugs |
| [16:11] | Andrew Linden: | I'm going to re-open it. |
| [16:11] | Techwolf Lupindo: | Sorry, only the oriangle reporter can open it. |
| [16:12] | Ardy Lay: | Andrew has a plasma cutter and a pry bar. He can get it opened. ;-) |
| [16:12] | Liisa Runo: | llSLPPFast(LINK_ROOT, [stuff..]); still fails? cant find the jira |
| [16:12] | Andrew Linden: | Last item... what I've been working on this week is a minor optimization of terrain collisions -- one of the top "hot spots" in our physics engine interface code. |
| [16:12] | Andrew Linden: | Not a big deal, but figured I'd mention it. It's what I've been doing. |
| [16:12] | Andrew Linden: | Ok. Table is now open for new topics. |
| [16:12] | Liisa Runo: | it is big deal |
| [16:12] | Sindra Aquila: | what are the improvements in land collision? |
| [16:13] | Kaine Lowenstark: | this means I wont be a foot in ground on flat terrain :D? |
| [16:13] | Andrew Linden: | Well, we were doing a lot of math to get the terrain height in collision events, and then not using the results of the math. |
| [16:13] | Andrew Linden: | So I'm cleaning up the code to make it much more efficient. |
| [16:13] | Sindra Aquila: | has it become faster reacting on slow sims? |
| [16:13] | Andrew Linden: | I don't think the effects of my work will be very measurable by most SL residents |
| [16:13] | Morgaine Dinova: | Every bit helps. |
| [16:13] | Sindra Aquila: | thanks |
| [16:14] | Andrew Linden: | but it was a performance hot spot that showed up in some testing we did. |
| [16:14] | Liisa Runo: | is it the magic invis terrain above visible one? |
| [16:14] | Kaine Lowenstark: | Whatever was in that last server update did crack on the sims Im working on |
| [16:14] | Andrew Linden: | No Liisa, the "magic inconsistency between the visible terrain and its collision shape" is caused by other stuff |
| [16:14] | Kaine Lowenstark: | Everyones coasting over 43 FPS |
| [16:14] | Pauline Darkfury: | Are you aware of SVC-6020, Andrew, it could be related to terrain collision height - size of collision sphere for sculpty prims and AV Z position on them (changed behaviour between 1.38 and 1.40) |
| [16:14] | Sculpted prim | AV collision (standing/walking) behaviour has changed between 1.38 and 1.40 |
| [16:14] | Techwolf Lupindo: | Andrew, have you done any work on the http textrue on the sim side? I got a few notes from my testing of it. |
| [16:15] | Andrew Linden: | in particular, we transmit the terrain with a "lossy algorithm" |
| [16:15] | Sindra Aquila: | has any been done on restitution of material in ralation to collision? |
| [16:15] | Andrew Linden: | basically it is wavelet appriximations of the true terrain shape |
| [16:15] | Andrew Linden: | so if you have a very steep transition in the heightfield |
| [16:15] | Morgaine Dinova: | Andrew: do you work with terrain at a level where you could be involved in the design of SL Terrain Next Gen? The SL terrain really is distressingly horrible, compared to everything else (especially water). The edges at high curvature, the stretched textures ... yuck. Needs version #2. |
| [16:15] | Andrew Linden: | you will see "ringing" effects there |
| [16:16] | Andrew Linden: | where some of the higher order terms are missing, but still macroscopically significant |
| [16:16] | Attica Bekkers: | new textures wld upset folks tho wldntthey? |
| [16:16] | Andrew Linden: | Yeah, I was talking with another LL dev this week about my ideas for fixing the terrain transmission pipeline |
| [16:16] | Attica Bekkers: | yay |
| [16:16] | Andrew Linden: | the main "puzzle" being how to do it without breaking too many builds |
| [16:17] | Kaine Lowenstark: | everythign upsets everyone, Im all better ^_^ |
| [16:17] | Andrew Linden: | since if we change the visible location of the heightfield (because we send it more accurately) then it will poke up above some floors (or will disappear below some others) where it did not b efore |
| [16:17] | Morgaine Dinova: | Progress always upsets somebody. You have to ignore them or your future is blocked. |
| [16:18] | Andrew Linden: | However, we came up with a scheme on how to do it. |
| [16:18] | Kaine Lowenstark: | Now could there not be blanket notices |
| [16:18] | Andrew Linden: | Which I can't remember now. But I could derive it again, I thihnk. |
| [16:18] | Kaine Lowenstark: | and have torley focus on terriforming skills in videos a few weeks before? |
| [16:18] | Youri Ashton: | there are different kinds of 'progress' morg, not all progress is good, neither is it always bad |
| [16:18] | Andrew Linden: | Oh yeah, this bit of news... |
| [16:18] | Pauline Darkfury: | How about an EM settable flag for enabling new terrain handling? |
| [16:19] | Andrew Linden: | I heard from another LL dev that one of the big lag spikes from rezzing MONO scripts is fixed in a test code branch. |
| [16:19] | Youri Ashton: | its the delecate matter of waging the good things and bad things before a decision can be made |
| [16:19] | Kaine Lowenstark: | Thats good news |
| [16:19] | Moon Metty: | mmmm good |
| [16:19] | Attica Bekkers: | OMG |
| [16:19] | Andrew Linden: | They are going to give it to the QA group to test soon. Might be in 1.44. |
| [16:19] | Attica Bekkers: | YAY |
| [16:19] | Andrew Linden: | That is "rumor" at this stage. It hasn't been tested. |
| [16:19] | Kaine Lowenstark: | Thats been one of the big bugs since Ive started coming to these OH's |
| [16:19] | Kaluura Boa: | Will there be new terrain textures one day? |
| [16:20] | Youri Ashton: | talking about lag Andrew, did you have the opertunity yet to look at the problem i gave you some time ago? |
| [16:20] | Kaine Lowenstark: | rumors are hope, hope is good |
| [16:20] | Attica Bekkers: | everytime someonelogsinto myhangout sim everyonensimis frozen, andthe people wot stop wearing aos, so thats great |
| [16:20] | Andrew Linden: | Someday Kaluura yes. But I'd expect new terrain textures and basic terrain features to be years away (at least two years). |
| [16:21] | Andrew Linden: | Youri, remind me of your problem please. |
| [16:21] | Kaluura Boa: | I'll be patient then... ::P |
| [16:21] | Andrew Linden: | You gave me a notecard? |
| [16:21] | Kaine Lowenstark: | So after 2012 =/ |
| [16:21] | Youri Ashton: | ye i did |
| [16:21] | Youri Ashton: | something with grasmere in the title |
| [16:21] | Sindra Aquila: | in relation to these big lagspikes from ppl coming in a sim, has any been done to restrict personal script usage and prim usage? |
| [16:21] | Youri Ashton: | Grasmere area problems |
| [16:22] | Andrew Linden: | I still have the notecard. |
| [16:22] | Morgaine Dinova: | Youri: I'm referring to "progress" in the fairly objective sense of "technically better" or "more advanced" or "more featureful" or "more visually impressive" or "more efficient". Some things are a bit subjective, but when the advance is sufficiently large, it becomes quite objective. |
| [16:22] | Youri Ashton: | earlier this week another person said the same things as I typed out |
| [16:22] | Andrew Linden: | The problems were about... getting stuck logged into a region. |
| [16:22] | Youri Ashton: | so the problem is still there it seems |
| [16:22] | Andrew Linden: | Not getting info sent to your viewer. |
| [16:23] | Youri Ashton: | so my thought about it being a server issue is correct then? |
| [16:23] | Moon Metty: | svc-4169? |
| [16:23] | t pay | group deeded items |
| [16:23] | Attica Bekkers: | if w stop people wearing aos they will leave our sim forver, freezing evrytime someone enters doesnt make them leave |
| [16:23] | Andrew Linden: | I have not been able to do any legwork on that issue yet. |
| [16:23] | Moon Metty: | hmmm, we 're getting very few new reports |
| [16:23] | Sindra Aquila: | i mean a thousand and one redundant resizer scripts |
| [16:23] | Andrew Linden: | Some of the problems might be related to a "avatar presence" bug that another LL dev is looking into |
| [16:24] | Moon Metty: | something must have changed |
| [16:24] | Andrew Linden: | but that is the only news/idea that I've got at the moment |
| [16:24] | Youri Ashton: | I been hearing there might be already a old jira about it somewhere |
| [16:24] | Youri Ashton: | last monday if im not mistaken |
| [16:24] | Youri Ashton: | so it could be a old problem sticking its head out again |
| [16:24] | Kaine Lowenstark: | That do not rezz inworld script should be ARable =/ |
| [16:25] | Andrew Linden: | There are a few performance fixes that are pending for 1.44. I suspect any LL dev would like to try Grasmere on that new code first, and then try to reproduce the problem. |
| [16:26] | Andrew Linden: | Because that code will have much more diagnostic capabilities too (internal diagnostics, not tools from the viewer UI) |
| [16:26] | Moon Metty: | ok |
| [16:26] | Moon Metty: | , |
| [16:27] | Youri Ashton: | I hope you did notice I was talking about my alt that has it on 4 viewers, however, I tried it on 3 other avi's, also on 4 different viewers, no problem |
| [16:27] | Youri Ashton: | so it can not be DNS related (or at least not on my side) |
| [16:27] | Andrew Linden: | Youri, you're saying the problems are very much avatar specific? |
| [16:27] | Youri Ashton: | I fear it is |
| [16:27] | Youri Ashton: | I believe it might be avitar specific on the server itself |
| [16:28] | Andrew Linden: | An intriguing clue, but I'm not sure what it means. |
| [16:28] | Youri Ashton: | there are just a handfull of residents that seem to have it |
| [16:28] | Youri Ashton: | xstorm had the same issues he said |
| [16:28] | Andrew Linden: | It doesn't matter what the avatar is wearing? You pulled off all attachments right? That is what the notecard said. |
| [16:28] | Youri Ashton: | as one of the more vermilliar names |
| [16:29] | Youri Ashton: | I tried having the full avi, I tried having a nakid avi |
| [16:29] | Youri Ashton: | no difference |
| [16:29] | Youri Ashton: | crash time varies between 3 minutes up to 1 hour |
| [16:29] | Latif Khalifa: | what's the ETA on 1.42 pilot roll? |
| [16:29] | Andrew Linden: | So, the only other thing that might matter would be... inventory contents, or ... groups maybe. I don't have many ideas. |
| [16:29] | Kaine Lowenstark: | I seem to be a bit lost, whats this? |
| [16:29] | Youri Ashton: | and for some odd reason, going over the sim line does NOT load the next sim |
| [16:30] | Andrew Linden: | 1.42 pilot roll was scheduled for Aug 18th I think, last time I looked. |
| [16:30] | Simon Linden: | Yep, still is |
| [16:30] | Youri Ashton: | ill get the pictures for those that do not know about it |
| [16:30] | Attica Bekkers: | ihave huge inventory |
| [16:30] | Attica Bekkers: | it doe snot effect my lag |
| [16:30] | Attica Bekkers: | or seem to effect others aroundme |
| [16:30] | Youri Ashton: | http://www.flickr.com/photos/42984343@N03/sets/72157624361935441/ |
| [16:30] | Attica Bekkers: | i canlagother by wearingthe wrong hair, tho |
| [16:30] | Youri Ashton: | all pictures in that folder are part of this bug |
| [16:31] | Latif Khalifa: | i've heard people report problems with llGetCenterOfMass() returning 0,0,0 after region crossing |
| [16:31] | Youri Ashton: | doesnt matter where you cross the sim |
| [16:31] | Kaine Lowenstark: | Ive had that bug during the SL 7B |
| [16:31] | Kaine Lowenstark: | no where else |
| [16:32] | Andrew Linden: | oh really Latif? Falcon will love to hear that center of mass is changing. |
| [16:32] | Simon Linden: | are they calling that right after the region crossing? |
| [16:32] | Latif Khalifa: | yes |
| [16:32] | Kaine Lowenstark: | ( the entering and finding your on " the only sim on mainland" |
| [16:32] | Andrew Linden: | Love as in "hate". But I'm sure he'll be able to fix it in about 1 hour. |
| [16:32] | Simon Linden: | ugh, that's another sync issue |
| [16:33] | Youri Ashton: | only thing i noticed is that i could fly off sim and hang around 1 sim over so to say without crashing or loading |
| [16:33] | Latif Khalifa: | i was wondering if it was similar to problem with libomv agents getting at 0,0,0 after region entry in 1.40 |
| [16:33] | Simon Linden: | I know what's going on ... 1.40 and H7 queue up changes and process them at a slightly different time |
| [16:33] | Youri Ashton: | it doesnt matter either what sim you use it seems |
| [16:33] | Andrew Linden: | Oh right. There is a bit more lash between the physics notion of the object's info and ours. |
| [16:33] | Simon Linden: | So before it would make your object, add the physics info, then run your scriopt |
| [16:33] | Youri Ashton: | i think its really avitar specific |
| [16:34] | Simon Linden: | Now it's making the object, running the script, updating physics |
| [16:34] | Latif Khalifa: | ah, i see |
| [16:34] | Andrew Linden: | right. |
| [16:34] | Youri Ashton: | hey arawn |
| [16:34] | Simon Linden: | That's the same thing that caused the libomv bots to end up a 0,0,0 |
| [16:34] | Andrew Linden: | oh really Simon? What bots were those? |
| [16:35] | Andrew Linden: | I was trying to help someone with a recent bot problem that sounds identical. |
| [16:35] | Simon Linden: | older libomv-based bots |
| [16:35] | Arawn Spitteler: | /me replaced SVC-93 with SVC-6070, and wonders what to do with SVC-22 and Sheep |
| [16:35] | Rotations are | borked for llSetLinkPrimitiveParams and Child Prims. SVC-93 |
| [16:35] | Vehicles crossing | region borders aren't always treated as vehicles and can get incorrectly returned if the destination parcel is no-entry or parcel-full |
| [16:35] | Youri Ashton: | first time i told you Andrew, was in this OH just before you went on vacation, you called it 'critical' at that time |
| [16:35] | Latif Khalifa: | yeah, i spoke with Ardy who said a first couple llGetCenterOfMass() calls after region crossing returned 0,0,0 and then later calls would work fine |
| [16:35] | Simon Linden: | it could be done on any client, if it sent an AgentUpdate and we got it in the same frame when the AV logged into the region ... if we processed that message before the AV was fully inserted inthe physics engine, it would still have a 0,0 position rather than the log in spot |
| [16:36] | Pauline Darkfury: | I can confirm that - I had 3 traffic camping bots who were camping a 0,0 mainland parcel of mine, without my consent or wanting any such things on my land. No amout of banning, TP home, orbit, etc would get rid of them, they just pinged right back |
| [16:36] | Ardy Lay: | Huh? Wasn't me that said that. :-) |
| [16:36] | Pauline Darkfury: | looked like they were constantly trying to get to 0,0,0 |
| [16:36] | Latif Khalifa: | ah, i must have mixed you up, sorry |
| [16:37] | Pauline Darkfury: | (a couple of ARs and a couple of live chats eventually got rid of them) |
| [16:37] | Moon Metty: | SVC-6027 |
| [16:37] | 4 breaks | libomv clients (and possibly gets corrupted in the process) |
| [16:37] | Latif Khalifa: | funny thing is i cannot reproduce the same thing on aditi 1.40 regions ;) |
| [16:38] | Attica Bekkers: | a lot of tps etc still go to 000 before destination |
| [16:38] | Latif Khalifa: | like ahern |
| [16:38] | Latif Khalifa: | must be slower hardware and timing issues that causes it not to repro on aditi |
| [16:39] | Simon Linden: | yeah, it's definitely a timing issue |
| [16:39] | Andrew Linden: | Ok, so the avatar arrives, and if you perform a certain operation too soon then your avatar ends up at (0,0,0)? |
| [16:39] | Youri Ashton: | I could test it on test grid if needed, my alt is showing lately odd behaviors anyways. perhaps it shows on test grid as well |
| [16:39] | Ardy Lay: | Latif, might try it in Lusk. We have people using Radegast there that still end up sulking in the corner even after updating. |
| [16:39] | Latif Khalifa: | Andrew, yes |
| [16:40] | Andrew Linden: | An operation that a regular user with the SL UI to contend with typically can not perform in time to reproduce. |
| [16:40] | Andrew Linden: | ok that's a good clue. |
| [16:40] | Latif Khalifa: | we sent AgentUpdate too soon for 1.40 |
| [16:40] | Youri Ashton: | ill be testing there tomorrow then |
| [16:41] | Latif Khalifa: | the trouble was, physics would move avatar to 0,0,0 irrespecitve of bans, access lists, etc etc ;) |
| [16:41] | Moon Metty: | hehe |
| [16:42] | Andrew Linden: | This just in our internal IRC channel: "another point for TechWolf! SVC-6073" |
| [16:42] | texture service | sends Content-Type: text/html instead of Content-Type: image/x-j2c |
| [16:42] | Techwolf Lupindo: | /me lol |
| [16:42] | Moon Metty: | :D |
| [16:42] | Techwolf Lupindo: | I got another one too. |
| [16:42] | Latif Khalifa: | the more important one was keep-alive being disabled |
| [16:42] | Attica Bekkers: | LOL |
| [16:42] | Latif Khalifa: | but i can sort of understnd why |
| [16:43] | Techwolf Lupindo: | I'me really SHOCK that the jiras got triaged and NOTICED!!! |
| [16:43] | Attica Bekkers: | yay triage:> |
| [16:43] | Kaluura Boa: | http://jira.secondlife.com/browse/SVC-6069 |
| [16:43] | Http GET | texture not using keep-alive on sim side. |
| [16:44] | Techwolf Lupindo: | Today, during my tweeking and testing of using a local squid proxy, found anther bug in the http texture sim server. |
| [16:45] | Morgaine Dinova: | The timing issues are worrying. How are events and state transitions handled on the sim? Are messages buffered up and then picked off the queue when state transition rules allow it, or are messages that are not valid in the current sim state rejected? |
| [16:45] | Ardy Lay: | Cascaded squids |
| [16:45] | Youri Ashton: | think it might be a wise thing to fix things first now before we adding new things, seems to be a lot of problems that can be called bad |
| [16:45] | Attica Bekkers: | I have a q if you have time will changes to physics be tested for cam views with cognatively different and sensitive people? |
| [16:46] | Andrew Linden: | messages are queued up and processed in order, there is no "state filter" that I know of |
| [16:46] | Attica Bekkers: | last big change me and every other middle aged womani knew got seasick |
| [16:46] | Youri Ashton: | sea sick in SL??? |
| [16:46] | Andrew Linden: | most of the messages are sanity checked |
| [16:46] | Attica Bekkers: | physics chnaged and it effected cam, ll got straight onit and it fixed fast |
| [16:46] | Andrew Linden: | I think there are some out-of-order checks for certain kinds of messages. |
| [16:46] | Arawn Spitteler: | Sanity Checked in SL? |
| [16:47] | Morgaine Dinova: | Processed in order of receipt, I take you to mean. So it's possible for an invalid state transition to occur unless the processing validates at every step. |
| [16:47] | Andrew Linden: | "most" ;-) |
| [16:47] | Latif Khalifa: | *cough* ViewerEffect *cough* :P |
| [16:47] | Youri Ashton: | sanity is long gone it seems, too much jira's getting closed by residents that 'think' that they can be closed without it being checked by any linden |
| [16:48] | Youri Ashton: | lots of bugs slip trough like that |
| [16:48] | Attica Bekkers: | they relaly did change how cam worked a little, and really did fix becaus eof "motion sick" folk |
| [16:48] | Youri Ashton: | think you have a fair point there attica |
| [16:49] | Youri Ashton: | isnt really concidered, but its a fair point to look at |
| [16:49] | Arawn Spitteler: | /me recalls Motion Sickness, and hopes QA does as well |
| [16:49] | Andrew Linden: | sigh... you're talking about the Havok4 camera oscilations? |
| [16:49] | Youri Ashton: | i rather meant in general andrew |
| [16:49] | Attica Bekkers: | chnages tpo hysics in progres snow so young cognatively fit folks may decide to tweak cam to be more real again, so im raising tests with physically challenged groups as apossible course of action |
| [16:50] | Moon Metty: | hmmmm SVC-2049 was one of the unclosed issues LOL |
| [16:50] | Camera zooms | rapidly in and out (yo-yo, pogos) with objects behind you |
| [16:50] | Andrew Linden: | Attica, could you elaborate on that statement? |
| [16:51] | Latif Khalifa: | Moon close it :) |
| [16:51] | Moon Metty: | now! |
| [16:51] | Moon Metty: | :) |
| [16:51] | Morgaine Dinova: | /me is really happy that English has a lot of redundancy, when reading Attica's lines ^_^ |
| [16:51] | Youri Ashton: | she is in im with me, ill try to make a corect text since she is dislectic as she tells me |
| [16:51] | Attica Bekkers: | is too |
| [16:51] | Morgaine Dinova: | Hehe |
| [16:52] | Youri Ashton: | "one of the lindens responded to a comment I made on a blog" |
| [16:52] | Youri Ashton: | "and they had changed the cam" |
| [16:52] | Ardy Lay: | I remember the camera oscillating violently around Christmas time. |
| [16:52] | Youri Ashton: | "they tweaked it back" |
| [16:52] | Andrew Linden: | This past Christmas? |
| [16:52] | Ardy Lay: | Yes. |
| [16:52] | Youri Ashton: | "and the remaining changes we just had to 'learn' " |
| [16:53] | Ardy Lay: | I moved some trees to keep people from complaining, to make them comfortable. |
| [16:53] | Latif Khalifa: | Ardy, depends on where you are. High up at about 3-4 km you start to get floating errors ;) |
| [16:53] | Attica Bekkers: | no 2008 |
| [16:53] | Ardy Lay: | We were at 61 meters |
| [16:53] | Ardy Lay: | The trees causing the issue were stacked sculpted prims set phantom. |
| [16:54] | Youri Ashton: | hmm.. having a little trouble with the next one, but so fr i understand is about teh mouselook looking more 'realistic' for a cam |
| [16:54] | Andrew Linden: | the camera code. There's a bad memory. We need to move ALL the camera control to the viewer. |
| [16:54] | Ardy Lay: | Hmm.. I may not have been here during the time you describe, Attica. |
| [16:54] | Youri Ashton: | "the last physical change made people feel motion sick is the thing" |
| [16:54] | Moon Metty: | what about scripted camera control? |
| [16:55] | Sindra Aquila: | i suppose you mean figurative, no really? |
| [16:55] | Youri Ashton: | "so if they want to tweak the cam, please get them to test it on their grannies" |
| [16:55] | Ardy Lay: | Unexpected camera motion can and does make people ill. |
| [16:55] | Andrew Linden: | the "last physical change" == Havok7 ? |
| [16:55] | Attica Bekkers: | more viewer cam control may be justthe thing. If thedefault is set for the motion sick prone not the young emen who dont throw up in mouselook:> |
| [16:55] | Youri Ashton: | "they need to be carefull of tweaking the cam" |
| [16:55] | Sindra Aquila: | you mean to tell me that ppl really can get so bad affected by a screen? |
| [16:56] | Attica Bekkers: | yes |
| [16:56] | Al Sonic: | It happens. |
| [16:56] | Latif Khalifa: | I have not noticed any camera changes with 1.40 |
| [16:56] | Attica Bekkers: | italways astounds youn folk |
| [16:56] | Youri Ashton: | okay, thats about all :p |
| [16:56] | Attica Bekkers: | no |
| [16:56] | Sindra Aquila: | huh? thgat is new to me, thanks |
| [16:56] | Attica Bekkers: | its been great, im just concerned for future |
| [16:56] | Ardy Lay: | Sindra, yes, and the sidebar in viewer 2 causing a point of view change does it to people as well. |
| [16:56] | Morgaine Dinova: | I've been fighting against unsolicited camera movements for some 4-5 years. Anyone want a "Leave My Camera Alone!" t-shirt? |
| [16:56] | Attica Bekkers: | thanks |
| [16:56] | Moon Metty: | Morg +1 |
| [16:56] | Latif Khalifa: | Morgaine, that's viewer thing |
| [16:57] | Moon Metty: | yes, a very annoying viewer thing :) |
| [16:57] | Simon Linden: | there's a pref for that sidebar shifting, fwiw |
| [16:57] | Morgaine Dinova: | Yep, it is, but Andrew mentioned moving some remainding camera stuff into the viewer |
| [16:57] | Simon Linden: | I can't stand it either |
| [16:57] | Sindra Aquila: | v2 ia unuable for games for several reasons any way |
| [16:57] | Ardy Lay: | I would say there are two classes of unwanted camera motions. Viewer action induced and simulator "object avoidance" induced. |
| [16:58] | Youri Ashton: | would it be a good suggestion to add in a on/off button for that? then people can enable the more 'realistic' cam on their own if they wish |
| [16:58] | Andrew Linden: | the server currently does some raytracing to figure out if your camera location will be unable to view your avatar |
| [16:58] | Youri Ashton: | both parties will be having what they want then |
| [16:58] | Andrew Linden: | if so, then it supplies a "plane" to the viewer which is a hint: don't put your camera behind this plane |
| [16:58] | Andrew Linden: | the viewer code can decide to pay attention to that plane or not |
| [16:58] | Moon Metty: | aha |
| [16:58] | Latif Khalifa: | yeah "keep me and my camera on the same side f the wall" thing |
| [16:58] | Attica Bekkers: | the current system is working great, just dont wantit changing to much withut some testing on cognatively disabled |
| [16:58] | Morgaine Dinova: | Aha, understood Andrew |
| [16:58] | Andrew Linden: | the oscillation happens when the server offers different planes very rapidly |
| [16:59] | Ardy Lay: | Andrew, sometimes that oscillates |
| [16:59] | Latif Khalifa: | that was mostly fixed in the recent versions of the sim |
| [16:59] | Andrew Linden: | what I want is for the viewer to comput its own "plane" |
| [16:59] | Attica Bekkers: | thats great Andrew |
| [16:59] | Andrew Linden: | so that the server doesn't need to comput it (less work for the server) |
| [16:59] | Sindra Aquila: | anyine else notice the big problem with v2 , the cluttering up of mouse screen with too many laines some at the wrong place besides? |
| [16:59] | Morgaine Dinova: | Good idea |
| [16:59] | Moon Metty: | that makes sense Andrew |
| [17:00] | Youri Ashton: | good suggestion andrew |
| [17:00] | Andrew Linden: | the viewer can't do the camera location until it has decent raytrace code |
| [17:00] | Andrew Linden: | so the project has been waiting for that |
| [17:00] | Moon Metty: | so we need Qarl |
| [17:00] | Morgaine Dinova: | :-) |
| [17:01] | Al Sonic: | hah. |
| [17:01] | Andrew Linden: | yeah, or we nee a decent phys engine on the viewer |
| [17:01] | Youri Ashton: | but wouldnt it be better to have texture for example saved to the HDD as well? if its both on the server as the HDD it will cost less work for the server, but also less to load, so speed goes up |
| [17:01] | Moon Metty: | :) |
| [17:01] | Morgaine Dinova: | Andrew: a closed source binary module is NOT an acceptable "decent physics engine" for the viewer, and there's going to be fireworks if Falcon tries to put that in. |
| [17:02] | Morgaine Dinova: | We had enough trouble getting FMOD removed. |
| [17:03] | Andrew Linden: | Yeah Morgaine. Falcon is set on that closed binary. |
| [17:03] | Sindra Aquila: | are ther any plans to change the now unusable mouselook screen of v2 into the old version of it? |
| [17:03] | Andrew Linden: | However, there would be hooks for an open source solution. Someone would have to write it. |
| [17:03] | Andrew Linden: | If we proceed down the path that Falcon has planned. |
| [17:04] | Kaluura Boa: | /me wonders a full physics engine in the viewer? What for? |
| [17:05] | Andrew Linden: | Well Kalurra for one, such a physics engine would provide efficient raytrace checks |
| [17:05] | Andrew Linden: | which could be used to compute camera location that does not occlude your avatar. |
| [17:06] | Arawn Spitteler: | Rubber Banding? |
| [17:06] | Arawn Spitteler: | Psui looked so cute |
| [17:07] | Moon Metty: | hmmm, rubber banding would still be there, i guess |
| [17:07] | Morgaine Dinova: | That closed binary will split the community in two, closed versus open source, and even if an open source version were programmed, it would not deliver the same physics results as the closed one --- there's plenty of experience with ODE behaviour in Opensim being very different from those of Havok here. So it will split the community in two. |
| [17:08] | Andrew Linden: | No, Falcon believes he can fix rubber banding if he has a decent phys engine on the viewer. |
| [17:08] | Andrew Linden: | That and a new communication protocol to the server. |
| [17:08] | Moon Metty: | but rubber banding occurs when the server and client lose contact |
| [17:08] | Moon Metty: | it could improve of course |
| [17:09] | Moon Metty: | but it can't be eliminated |
| [17:09] | Ardy Lay: | Would have to eliminate the halting of physics in simulators. |
| [17:09] | Moon Metty: | unless the server and client always keep contact |
| [17:09] | Simon Linden: | I have to run and juggle some code ... thanks everyone for coming |
| [17:09] | Andrew Linden: | A slight modification of that statement Moon: it happens when client and server lose contact AND when their simulations diverge |
| [17:09] | Simon Linden: | See you next time |
| [17:09] | Morgaine Dinova: | Cyu Simon, have a good weekend |
| [17:09] | Moon Metty: | good luck Simon :) |
| [17:09] | Andrew Linden: | if the veiwer's simulation is more correct, then you have less divergence |
| [17:09] | Simon Linden: | Bye everyone |
| [17:09] | Moon Metty: | yes Andrew |
| [17:10] | Moon Metty: | if you move in a straight predictable motion |
| [17:10] | Moon Metty: | but when you turn, it changes everything |
| [17:11] | Andrew Linden: | if you have the same data and phys engine then you can also do well in some non-linear motion |
| [17:11] | Ardy Lay: | four axis and velocity motion delta prediction, whee |
| [17:11] | Andrew Linden: | but the scheme for doing that is a bit tricky |
| [17:12] | Andrew Linden: | you can do some reverse simulation of old control data to recreate the avatar's motion, even in extreme lag (much of the time) |
| [17:12] | Andrew Linden: | but there is a bunch of work to do to make that possible |
| [17:12] | Andrew Linden: | I've always been skeptical of going that route |
| [17:13] | Andrew Linden: | and have preferred to just try to rely on reduced lag of the client-server connection |
| [17:13] | Moon Metty: | yes |
| [17:13] | Ardy Lay: | Dig up an apollo guidance computer simulator. ;-) |
| [17:13] | Moon Metty: | the more successful updates the better |
| [17:13] | Andrew Linden: | however Falcon has a plan, and we have struggled with the server lag forever -- |
| [17:13] | Andrew Linden: | perhaps my preferred mode just won't work in practice |
| [17:14] | Andrew Linden: | it certainly doesn't work for distant clients (high ping time) who like the immediate response of FPS avatar controls |
| [17:14] | Andrew Linden: | FPS = first person shooter |
| [17:14] | Moon Metty: | ok |
| [17:14] | Sindra Aquila: | very important sl par, lol |
| [17:15] | Moon Metty: | it's possible to do an interpolation after rubber banding |
| [17:15] | Andrew Linden: | Well, I'm going to head out. Have a good weekend everyone. |
| [17:15] | Andrew Linden: | Thanks for coming. |
| [17:15] | Sindra Aquila: | thanks andrew |
| [17:15] | Arawn Spitteler: | thx 4 staimg |
| [17:15] | Moon Metty: | from where the client thinks it should be to where the server thinks it should be |
| [17:15] | Moon Metty: | thanks Andrew :) |
| [17:15] | Ardy Lay: | Thanks Andrew |
| [17:16] | Moon Metty: | have a good weekend |
| [17:16] | Sindra Aquila: | bye everyone |
| [17:16] | Youri Ashton: | pokes kitty |
| [17:16] | Moon Metty: | hehe |
| [17:16] | Morgaine Dinova: | So, Falcon is out of control. Great. |
| [17:16] | Youri Ashton: | not that kitty moon ;) |
| [17:16] | Youri Ashton: | haha |
| [17:17] | Moon Metty: | lol |
| [17:17] | Moon Metty: | why would it be closed? |
| [17:17] | Morgaine Dinova: | Because Falcon comes from Havok, and that's how his mind works. |
| [17:17] | Moon Metty: | hmmm |
| [17:18] | Morgaine Dinova: | He's clearly not in tune with the opening of the client, and slow removal of closed stuff. |
| [17:18] | Moon Metty: | well the way i see it |
| [17:19] | Attica Bekkers: | thatwas a good meeting |
| [17:19] | Moon Metty: | it's possible (with great effort) to code it |
| [17:19] | Attica Bekkers: | thankyou fro trying to understand me |
| [17:19] | Youri Ashton: | no worries :) |
| [17:19] | Morgaine Dinova: | Yep, very good meeting. Just an unfortunate revelation for open source. |
| [17:19] | Moon Metty: | but you can't repilcate the behavious of closed code, never |
| [17:19] | Latif Khalifa: | /me waves to everyone "have fun guys" |
| [17:19] | Attica Bekkers: | be well latif |
| [17:19] | Moon Metty: | see you Latif :) |
| [17:19] | Morgaine Dinova: | In physics, we know from experience that you can't duplicate it, aye. |
| [17:20] | Moon Metty: | right, in physics |
| [17:20] | Bronson Blackadder: | poking me |
| [17:20] | Bronson Blackadder: | lol |
| [17:21] | Ardy Lay: | I would imagine it's more difficult to emulate an inaccurate simulation than to create a reasonably accurate one. |
| [17:21] | Youri Ashton: | pokes kitty again |
| [17:21] | Attica Bekkers: | loeths ultracat series bronson |
| [17:21] | Bronson Blackadder: | lol |
| [17:21] | Moon Metty: | lol yes Ardy |
| [17:21] | Attica Bekkers: | i have two |
| [17:21] | Morgaine Dinova: | Probably, Ardy |
Generated with SLog Wikifier