User:Andrew Linden/Office Hours/2008 06 10

Transcript of Andrew Linden's office hours:

[11:03]	Teeple Linden:	Ah! Got it :D
[11:03]	Teravus Ousley:	And he arrives.
[11:04]	Andrew Linden:	sorry I'm late
[11:04]	Teravus Ousley:	Thanks for coming. :D
[11:04]	Rex Cronon:	HI ANDREW
[11:04]	Andrew Linden:	I was helping another LL dev figure out some problem.
[11:04]	Johan Laurasia:	Hiya Andrew
[11:04]	Johan Laurasia:	http://jira.secondlife.com/browse/MISC-1261
[11:04]	Rex Cronon:	oops. caps was on:)
[11:04]	Johan Laurasia:	there's the URL to those interested
[11:05]	Andrew Linden:	I don't have much for announcments... other than my own lack of progress on stuff.
[11:05]	Andrew Linden:	I spent some time fixing emergency problems with the 1.21 and 1.22 deploys.
[11:05]	Simon Linden:	Hey, that looks like a good bug. Thanks - hopefully it will be easy to reproduce ... if so, it will be good to fix
[11:06]	Andrew Linden:	And am still working on a few ... some VolumeDetect vs sitting agents on phantom vs non-phantom seats problems.
[11:06]	Johan Laurasia:	yeah, it's not too tough
[11:06]	Teravus Ousley:	hehe, finding a good sit spot... is actually more challenging then it appears.
[11:07]	Andrew Linden:	too bad Kelly isn't still working on Havok4. I'd love to terrorize his day by assigning that bug to him.
[11:07]	Arawn Spitteler has found an undocumented benefit, if sitting logic, and hopes it won't be "fixed" in the future: Is there any limit to SLPP?
[11:07]	Andrew Linden:	The MISC-1261 bug.
[11:07]	Andrew Linden:	? SLPP?
[11:07]	Johan Laurasia:	um... idk?
[11:08]	Johan Laurasia:	what are you asking?
[11:08]	Arawn Spitteler:	llSetLinkPrimitiveParams(integer, list) by which people can be moved about in sailboats. I'm using it in a TP, that exceeds 300 meters
[11:08]	Andrew Linden:	Ah
[11:08]	Simon Linden:	Andrew, I'll send him a link for that bug and let you know what the reaction looks like
[11:09]	Arawn Spitteler:	It's great for Spirit City, where we have classrooms otherwise reached by warpPos
[11:09]	Andrew Linden:	I dunno Simon, I think he feels safe now that he's working on some new feature.
[11:09]	Andrew Linden:	I don't think the response will be quite genuine.
[11:09]	Teravus Ousley:	:D
[11:10]	Andrew Linden:	Anyway... I guess the only new bit of news is that I'm almost done fixing the region-wide prim limit problem... we aren't yet really enforcing regionwide prim limits in some cases, but we will when my stuff goes out.
[11:11]	Rex Cronon:	how is that going to work andrew?
[11:11]	Andrew Linden:	Anybody have topics to discuss?
[11:11]	Andrew Linden:	The way it used to work Rex was that if the region became over-full then no one would be able to rez stuff until it fell below the prim limit.
[11:12]	Arawn Spitteler:	I was surprised to find Flight Feather/X-Flight is no longer needed in Spirit City
[11:12]	Creem Pye:	what? did you already remove the flight height limit?
[11:12]	Andrew Linden:	That was pre-havok4. I ripped out that check in Havok4 to clean up some code... I mistakenly believed that the parcel prim accounting was tight enough to not need the region-wide checks anymore.
[11:12]	Andrew Linden:	I was mistaken.
[11:12]	Arawn Spitteler:	Not for all sims, though
[11:12]	Johan Laurasia:	I've got a question concerning mini sims
[11:13]	Andrew Linden:	What is a "mini sim"?
[11:13]	Johan Laurasia:	um... low prim
[11:13]	Gaius Goodliffe:	Openspaces.
[11:13]	Teravus Ousley:	void
[11:13]	Johan Laurasia:	openspaces...
[11:13]	Andrew Linden:	ok, aka "void regions" or... "open space regions" I think they are called.
[11:13]	Johan Laurasia:	a friend has 271 prims rezzed, but 'simulator primitive usage' reads 338, but the list of individuals shows correct.
[11:13]	Gaius Goodliffe:	Openspaces(TM) is the proper Second Life(R) term. ;)
[11:14]	Andrew Linden:	"lowprim sims" might be a more appropriate name.
[11:14]	Gaius Goodliffe nods.
[11:14]	Johan Laurasia:	well, the question being, is 'simulator primitive usage' showing incorrect values?
[11:14]	Andrew Linden:	Johan, I'm confused by the question... ok so someone owns a parcel with a 338 prim limit, and is using 271?
[11:15]	Johan Laurasia:	oh, no, she has about 1/3rd the sim
[11:15]	Johan Laurasia:	she has (in reality) 271 prims rezzed
[11:15]	Johan Laurasia:	yet 'sim prim usage' shows 338 used
[11:15]	Gaius Goodliffe:	Ghost prims...
[11:15]	Johan Laurasia:	the individual list was showing correct though
[11:15]	Arawn Spitteler:	Ghost Prims shouldn't be counted, should they?
[11:15]	Gaius Goodliffe:	Try rebooting them sim, see if they disappear.
[11:15]	Andrew Linden:	where is the "simulator primitive usage" stat coming from? The CTRL+SHIFT+1 stats?
[11:16]	Gaius Goodliffe:	Ghost prims shouldn't exist to begin with. But yes, when they are there, they do count.
[11:16]	Johan Laurasia:	standard about land window
[11:16]	Johan Laurasia:	about land.. objects tab
[11:16]	Rex Cronon:	can u see them? can u edit them?
[11:16]	Andrew Linden:	"ghost prims" would not show up... those are prims that the main simulator code doesn't know about but are still in the physics engine.
[11:17]	Gaius Goodliffe:	You can't see or edit the ghost prims, but you collide with them.
[11:17]	Johan Laurasia:	so a sim reboot should take care of it
[11:17]	Johan Laurasia:	there doesnt appear to be any ghost prims about
[11:17]	Andrew Linden:	btw, I haven't seen a real instance of a ghost prim since Havok4 launched... not that they don't exist, but they haven't been reproduced in my presence yet.
[11:17]	Johan Laurasia:	yeah, I've not run into one since 4
[11:17]	Arawn Spitteler:	Just had a ghost car at Lummerland
[11:17]	Gaius Goodliffe:	I've seen them personally once, and got one report of them.
[11:17]	Andrew Linden:	I was just hearing about a big fat ghost prim report that was actually a big flexi that sagged well below its colliding shape.
[11:17]	Rex Cronon:	i dont' think that i ever run into one
[11:17]	Creem Pye:	so "primitives on parcel" is the lower number?
[11:17]	Aeron Kohime:	Hello everyone.
[11:18]	Rex Cronon:	hi aeron
[11:18]	Creem Pye:	the difference might be from prims on other parcels, unless i'm misunderstanding
[11:18]	Johan Laurasia:	well, the individual list showed only she had prims rezzed on her lot
[11:18]	Andrew Linden:	Ok Johan, so the region-wide prim count doesn't distinguish between temp-on-rez and normal prims, I don't think.
[11:18]	Johan Laurasia:	she had just obtained the lot, and was sure to clear it first before starting her build
[11:18]	Andrew Linden:	Also... I wonder if it counts attachment prims...
[11:18]	Johan Laurasia:	hmm, could be
[11:19]	Andrew Linden:	It depends very much on which stat you are quoting, so where did you get it?
[11:19]	Johan Laurasia:	wouldnt that cause issues? some attachements have hundreds of prims
[11:19]	Teravus Ousley:	havok1, I'm pretty sure, didn't count them on any dialog/stat box.
[11:19]	Teravus Ousley:	(attachment prim)
[11:19]	Johan Laurasia:	about land... object tab... 'simulator primitive usage' was showing 338
[11:19]	Andrew Linden:	ok good to know Teravus
[11:19]	Johan Laurasia:	in reality she had 271 rezzed, and the individuals list showed only she had prims rezzed, and only 271
[11:19]	Creem Pye:	heh an avatar can have about 15k attached prims, so I'm sure simulator prim usage would get really high in a sim with a nightclub or something
[11:20]	Aeron Kohime:	The max not counting hud would be about 7650 prims
[11:20]	Johan Laurasia:	I could do some testing regard attached prims
[11:20]	Andrew Linden:	ah ok... that stat is total region prims... for all parcels
[11:21]	Johan Laurasia:	yep, she has just the 1 parcel
[11:21]	Andrew Linden:	so if she owns 1/3 the region... the missing 60+ prims are probably on a different parcel
[11:21]	Gaius Goodliffe:	9690 if HUD objects count :)
[11:21]	Johan Laurasia:	hmm, possible... I'll check into that too
[11:21]	Creem Pye:	oops guess I was off a bit :<
[11:21]	Johan Laurasia:	what was your question concerning MISC - 1261?
[11:21]	Sheet Spotter:	The "Simulator primitive usage" does not include attachments. I confirmed that by removing my 200+ prim hair. :-)
[11:22]	Teravus Ousley:	:D
[11:22]	Andrew Linden:	thanks for shaving one for the team Sheet
[11:22]	Johan Laurasia:	lol
[11:23]	Andrew Linden:	So was bouncing my latest great feature idea off of a fellow linden last week...
[11:23]	Andrew Linden:	and he just blew it off and said, "Do you want to hear my favorite feature idea?"
[11:23]	Arawn Spitteler:	Try filing a new features Jira?
[11:24]	Andrew Linden:	turns out, his "favorite feature" is a more responsive avatar control... runs immediatly when you press forward, stops when you stop.
[11:24]	Arawn Spitteler:	Is that doable?
[11:24]	Aeron Kohime:	they already have double-tap
[11:24]	Teravus Ousley:	network latency makes it tricky
[11:24]	Creem Pye:	hm that might feel a bit unnatural if the avatar accelerates too quickly
[11:24]	Andrew Linden:	However... he didn't actually say that... what he said he wanted was "client-side avatar prediction"
[11:25]	Arawn Spitteler:	Yes, that would be doable, with Client Side Physics
[11:25]	Johan Laurasia:	gee havent seen that since dial up doom days
[11:25]	Andrew Linden:	Well... some people have high ping-times... but it isn't all network latency.
[11:25]	Gaius Goodliffe:	I think we discussed this the other day after you had left...
[11:25]	Rex Cronon:	so there will be no more rubberbanding?
[11:25]	Andrew Linden:	Ok everyone... here is an exercise... state your pingtimes here:
[11:25]	Creem Pye:	there's some sort of avatar prediction with dead reckoning, right? when you get lagged, your avatar will continue along its velocity vector...
[11:25]	Andrew Linden:	mine is 95 msec
[11:25]	Johan Laurasia:	how do u turn that on again?
[11:25]	Creem Pye:	84ms
[11:25]	Holger Gilruth:	191 from germany
[11:25]	Simon Linden:	I've wanted to put a cap on the predicted motion - stop it locally if you press space bar, or taper it off after a while. It's odd when the network blocks and everyone starts walking off into the sunset
[11:26]	Andrew Linden:	CTRL+SHIFT+1 , then open oup the "Basic" tab at the top
[11:26]	Teravus Ousley:	157 msec from CT, USA
[11:26]	Gaius Goodliffe:	Simon: Yes!
[11:26]	Andrew Linden:	ping is... fourth item down
[11:26]	Aeron Kohime:	Speaking of this, the server side seems to be confused sometimes, I used to sometimes fly right through walls, until I decided to walk instead.
[11:26]	Arawn Spitteler:	426 from Bridgeport Ct
[11:26]	Aeron Kohime:	120 msec
[11:26]	Johan Laurasia:	95 msec from Lorain, Ohio
[11:26]	Sheet Spotter:	300 msec from Canada, eh.
[11:26]	Holger Gilruth:	will fastpath help to speed up?
[11:26]	Aeron Kohime:	from northern michigan
[11:26]	Rex Cronon:	how did u get your pingtime?
[11:26]	Rex Cronon:	oh. found it
[11:27]	Holger Gilruth:	Rex STRG+SHIFT+1
[11:27]	Gaius Goodliffe:	118 ms in MN
[11:27]	Aeron Kohime:	84 now, its all over the place, 444 msec
[11:27]	Simon Linden:	114 ms in California
[11:27]	Rex Cronon:	139
[11:27]	Aeron Kohime:	477... back down again 282, etc
[11:27]	Arawn Spitteler's is holding steady
[11:27]	Andrew Linden:	ok... so 100+ at the small side... up to half a second
[11:27]	Holger Gilruth:	sty between 181 and 240ms at 45 fps
[11:27]	Johan Laurasia:	I'm floating between 90 and 90
[11:27]	Johan Laurasia:	99*
[11:27]	Rex Cronon:	but i am on a wireless network
[11:27]	Aeron Kohime:	I would between say 100 and 500 here
[11:28]	Gaius Goodliffe:	Yeah, I usually hang around the 90's, but I'm also WiFi today...
[11:28]	Arawn Spitteler:	Slow response can cause motion sickness, with camera smoothing
[11:28]	Johan Laurasia:	usb connection direct to cable modem here, no router
[11:28]	Creem Pye:	hmm 500ms wouldn't be very pleasant if it stays at that =\
[11:28]	Sheet Spotter thinks Andrew just discovered a new party game: ping time poker.
[11:28]	Andrew Linden:	so, the SL client sends an "AgentUpdate" message with keyboard controls about 10 times a second... that is what has your avatar control flags in it.
[11:28]	Johan Laurasia:	lol
[11:28]	Holger Gilruth:	sometimes in high traffic areas (40 avatars) it suddenly go up to 2000 at me
[11:29]	Andrew Linden:	Your avatar won't start/stop/change faster than that... so in the worst-case for the lowest ping time...
[11:29]	Creem Pye:	I wonder how much apparent latency an AO might add, if it's manually polling for your keyboard controls and changing your animations based on that
[11:30]	Teravus Ousley:	that happens server side.. the AO, script..
[11:30]	Andrew Linden:	that would be 100 (freq) + 100 (ping) + 22 (sim fps) = 222 msec for the fastest server-only response
[11:30]	Gaius Goodliffe wonders how much more responsive things would be if we had AO events in LSL instead of polling.
[11:30]	Arawn Spitteler:	An AO can be fed keyboard, by control event, but they normally poll what the avatar is doing
[11:30]	Creem Pye:	AO events would be nice, but I'd rather have non-LSL custom animations for an avatar =)
[11:30]	Andrew Linden:	client-side prediction is necessary for long-ping FPS games
[11:30]	Creem Pye:	I guess the animations would be a subset of avatar appearance, then...
[11:31]	Andrew Linden:	I don't think AO's change the avatar motion response times.
[11:31]	Aeron Kohime:	Yeah, but it makes SL less open source... atleast if its havok.
[11:31]	Andrew Linden:	Not for the true physical change of position of your avatar's collision body.
[11:31]	Teravus Ousley:	Well, the other option is interpolating between the server position and the expected client position
[11:31]	Teravus Ousley:	.. once the server responds.
[11:32]	Andrew Linden:	btw, the SL avatar has a built-in timescale for starting and stopping motion... so the lag between the key press and the avatar's motion is not all network pingtime
[11:32]	Teravus Ousley:	.. it would be less accurate then running client side physics.. but you'd get going more. You'd have to anticipate what the PID controller would do though.
[11:32]	Creem Pye:	right, but people might not recognize the motion state change until the animation changes (like if I am walking and physically stop moving, I would still have the illusion of walking before my "walking" animation is stopped)
[11:33]	Andrew Linden:	oh yeah... assuming you've got 30 fps on your client... there is an additional 33 msec before you actually see your avatar moving... so more like 255 msec in the best normal-case
[11:33]	Arawn Spitteler tells Client: Go there. Client tells Server: I'm am going there, with such haste. Server tells Clients: these are the places, motions and collisions, smooth accordingly
[11:33]	Andrew Linden:	or 1/4 of a second for server-side-only updates (no client-side prediction)
[11:34]	Holger Gilruth:	Can i ask a little question at a full sim, with 14800 used prims and 30 avatars, is there a physic of 1.2 good or bad
[11:35]	Creem Pye:	physics time of 1.2ms? that's pretty good
[11:35]	Teravus Ousley:	1.2 time dilation sounds like it's trying to catch up.
[11:35]	Teravus Ousley:	.. 1.2 ms isn't bad at all :D
[11:35]	Holger Gilruth:	yes physicstime and scripttime is about 8 to 12ms
[11:36]	Andrew Linden:	uh... any TD > 1.0 would seem like an error to me
[11:36]	Andrew Linden:	but 1.2 msec in the physics step would be "good"
[11:36]	Holger Gilruth:	ive searched about 4 weeks to come from 5 to 1.3
[11:36]	Creem Pye:	a total time of <22.2ms is what you need for a perfect time dilation, Holger
[11:36]	Holger Gilruth:	its hard because you cant search directly for physic objects
[11:36]	Teravus Ousley:	:D yes, I recall working on simstats in opensimulator.. when ever it was trying to catch up, the time dilation would go higher then 1.0. :D
[11:37]	Andrew Linden:	i was thinking out loud about the avatar control lag and ping time because I think he avatar control could be made much tighter via server-only changes...
[11:37]	Andrew Linden:	at best (for normal case) the lag would be 1/4 second
[11:38]	Andrew Linden:	but it is more now... mostly because the there is built-in ramp-up of the avatar velocity
[11:38]	Holger Gilruth:	yes at high traffic hours (45 avas) we had normaly when nobody reez physic a dilletation about 0.8 and a physic of 1.6
[11:38]	Sheet Spotter:	I just sent my alt to a Class 5 sim with 66 agents. The "Sim Time (Physics)" was 0.4.
[11:39]	Andrew Linden:	in fact, there were times in the Havok4 beta (or pre-beta) where the avatar would start moving and would stop too fast... the testers would report the responsiveness as a bug
[11:39]	Holger Gilruth:	lol
[11:39]	Creem Pye:	yeah, I do think it would look weird if avatars accelerated instantly
[11:39]	Creem Pye:	at least that's not how we move in RL =P
[11:39]	Andrew Linden:	right creem, I suspect that many people would object if the avatar control were made more instantaneous
[11:39]	Johan Laurasia:	yeah, I really dont see a problem with avie movement
[11:40]	Johan Laurasia:	well, if someone wanted that, it's easy enough to script
[11:40]	Rex Cronon:	could there be new lsl functions that would alow to change the response time?
[11:40]	Arawn Spitteler:	Walking is hte art of falling, and catching ourselves, and standing requires a less dynamic ballance, so changing that ballance begins the first step
[11:40]	Creem Pye:	would there be a way to minimize control latency in other ways, by say having agent updates generate an interrupt?
[11:40]	Andrew Linden:	many people DO see a problem with avatar responsiveness... but so far most of them were FPS afficionados as far as I can tell (FPS = first person shooter)
[11:41]	Aeron Kohime:	Is it changable per avatar/onject in the physics engine?
[11:41]	Johan Laurasia:	yeah, I suppose if you're doing FPS it's a bit laggy
[11:41]	Teravus Ousley:	well there's a difference between 0-2mph in 0ms when you press the button.. then a slow ramp up to 2mph when you press the button..
[11:41]	Andrew Linden:	Rex, I don't even want to think about modifying LSL at this stage...
[11:41]	Arawn Spitteler:	LSL options are a Mono feature.
[11:41]	Andrew Linden:	even the smallest feature-add to the LSL engine is blocked by the MONO project
[11:41]	Teravus Ousley:	if you start moving immediately albeit, slowly accellerating.. .
[11:41]	Rex Cronon:	i was thinking that some flight assists might benefit from it
[11:42]	Andrew Linden:	and all of the work Babbage Linden has done that is not yet merged/deployed
[11:42]	Johan Laurasia:	yeah, it would be silly to create new LSL functions right now until mono is up and running solid
[11:42]	Johan Laurasia:	just askin for trouble if ya do
[11:42]	Creem Pye:	oh Rex, you could try using llmineventdelay() to increase the latency of a script
[11:42]	Holger Gilruth:	is mono only interesting for scripter or for user too
[11:42]	Holger Gilruth:	or sim owner
[11:43]	Andrew Linden:	The MONO project should make scripts run faster
[11:43]	Creem Pye:	(which would also make the script much lighter in terms of script time)
[11:43]	Rex Cronon:	thanks. i will try that
[11:43]	Gaius Goodliffe:	Mono is interesting to anyone with a high enough Geek Quotient.
[11:43]	Aeron Kohime:	Can I touch on a topic, it has to do with the physics engine, do avatars still coast through objects when its laggy?
[11:43]	Andrew Linden:	and may allow some projects that used to rely on multiple scripts to collapse them into one.
[11:43]	Johan Laurasia = high GQ
[11:43]	Teravus Ousley:	:D yes. key is faster.
[11:43]	Andrew Linden:	That is, some scripters would try to get around LSL performance problems by running multiple scripts in parallel
[11:43]	Johan Laurasia:	yep, 64k script size :) woot
[11:44]	Andrew Linden:	but MONO should allow (in some cases) the same performance (or faster) with a single script
[11:44]	Aeron Kohime:	I think most people run multiple scripts to get around mandatory sleep time in functions.
[11:44]	Gaius Goodliffe nods.
[11:44]	Creem Pye:	but there are still issues with functions that have a built-in 200ms delay, etc
[11:45]	Andrew Linden:	Aeron, the coasting through objects (and the snap-back) is also called "rubber banding" and is a lag + client-side extrapolation problem.
[11:45]	Aeron Kohime:	Ah but I don't snap back I go through a wall and end up there when everything settles
[11:45]	Andrew Linden:	oh... hrm...
[11:45]	Andrew Linden:	Aeron, is this when crossing region boundaries? Or even in the middle of a region?
[11:45]	Teravus Ousley:	bullet velocity through a wall scenario?
[11:45]	Creem Pye:	does that coincide with a sim crossing, Aeron?
[11:46]	Aeron Kohime:	middle of a region
[11:46]	Andrew Linden:	At slow speeds? or high?
[11:46]	Aeron Kohime:	mostly flying, rarely ever if walking
[11:46]	Aeron Kohime:	I don't normally run so I can't tell there
[11:46]	Andrew Linden:	Hrm... I haven't seen that happen yet, but not impossible.
[11:46]	Andrew Linden:	Do you have a location where it can happen regularly?
[11:46]	Andrew Linden:	Are you using flight assists?
[11:47]	Aeron Kohime:	most the time
[11:47]	Gaius Goodliffe:	I have a high-speed flight assist, so I fly through walls on a regular basis.
[11:47]	Andrew Linden:	What that "yes I use flight assists most of the time" ?
[11:47]	Aeron Kohime:	I can try it without and get back to you, but it doesn't boost unless I am above the physical flight zone
[11:47]	Johan Laurasia:	lol, sure do
[11:47]	Johan Laurasia:	never take it off... lol
[11:47]	Sheet Spotter:	What is your "Draw Distance"? Your client might not have the objects loaded yet.
[11:47]	Andrew Linden:	Ah hrm... does the flight assist push forward too?
[11:47]	Johan Laurasia:	'dont leave home without it'
[11:48]	Johan Laurasia:	lol
[11:48]	Aeron Kohime:	yes I use flight assists all the time
[11:48]	Holger Gilruth:	512 is my distance
[11:48]	Aeron Kohime:	I'm not sure
[11:48]	Aeron Kohime:	its the multi-gadget
[11:48]	Creem Pye:	;'boost' sounds like it might be tunneling your av through walls
[11:48]	Andrew Linden:	and what mode do the flight assists use? llMoveToTarget() or llApplyImpulse() or what?
[11:48]	Gaius Goodliffe:	Sheet: Low draw distance would cause the opposite problem
[11:48]	Teravus Ousley:	(server side collisions don't rely on draw distance)
[11:48]	Gaius Goodliffe:	Oh speaking of collision problems...
[11:49]	Andrew Linden:	yes the flight assist might help tunnelling. I'd have to examine the script to form a better opinion.
[11:49]	Aeron Kohime:	I'll try it without and tell you on thursday if I happen to be awake when that meeting happens
[11:49]	Andrew Linden:	ok thanks Aeron.
[11:49]	Gaius Goodliffe:	Is anyone else seeing prims where you only receive collision events reliably when you collide with particular faces, but don't receieve them or receive them very rarelly when colliding with other particular faces on the same prim?
[11:50]	Aeron Kohime:	I do know that physical object's go right through static prims when thier velocity is high (80+)
[11:50]	Andrew Linden:	is there anything the distinguishes the two types of faces Gaius?
[11:50]	Rex Cronon:	i think they go through when is > 150
[11:50]	Gaius Goodliffe:	Nope. And the problem disappears if you rotate the prim.
[11:50]	Creem Pye:	the minimum speed of tunneling depends on the thickness of the wall
[11:50]	Andrew Linden:	that is... does the effect differ across the faces of a box?
[11:51]	Andrew Linden:	or does it differ on the faces of a semi-sphere?
[11:51]	Gaius Goodliffe:	This is on a box.
[11:51]	Andrew Linden:	ok that sounds odd
[11:51]	Andrew Linden:	perhaps you could file a jira on it with a demo script that shows the problem
[11:52]	Gaius Goodliffe:	I have a set of eight doors on my observatory -- all identical (same scripts, and prims were created by simply duplicating the existing doors). Two of them have the problem, six don't.
[11:52]	Johan Laurasia:	speaking of jira, Andew, you asked a question concerning MISC - 1261, but I didnt quite understand it
[11:52]	Johan Laurasia:	my friend is the one who found/posted that
[11:52]	Gaius Goodliffe:	I'll file a JIRA. I'll see if I can't come up with a simpler repro too...
[11:53]	Johan Laurasia:	I have the offending object, I can give you a copy of it if you want
[11:53]	Andrew Linden:	Thanks Gaius, that really is the best way to file the info about the bug.
[11:53]	Simon Linden:	Johan - can you give me one? I grabbed teh bug
[11:53]	Johan Laurasia:	sure
[11:53]	Andrew Linden:	Johan, I can't remember the question I asked about MISC-1261
[11:54]	Johan Laurasia gave you 3MoreCubes.
[11:54]	Johan Laurasia:	u said SSPP? not sure what that meant
[11:54]	Andrew Linden:	Oh, someone used that acronym and I didn't know what it meant.
[11:54]	Johan Laurasia:	ah, me either
[11:54]	Andrew Linden:	but I think Arawn answered it...
[11:55]	Johan Laurasia:	well, the guy who found that is a newb, only a week or so old
[11:55]	Johan Laurasia:	and he found that, and IM'd me about it
[11:55]	Andrew Linden:	llSetLinkPrimParams = SLPP
[11:55]	Johan Laurasia:	I wasnt sure he was crashing the sim, just maybe the client
[11:56]	Johan Laurasia:	so we tested it from one sim over, and it definately brings the sim down
[11:56]	Andrew Linden:	Johan, when was that bug reported... ?
[11:56]	Andrew Linden looks
[11:56]	Johan Laurasia:	last saturday
[11:56]	Johan Laurasia:	it's an odd set to create the crash, but he's a newb
[11:56]	Andrew Linden:	ah... I fixed a crash mode that looks like it might be related to that... but that fix went out in 1.22.2
[11:57]	Arawn Spitteler:	The collisions aren't even physical, when the overlapping is unlinked, so llTargetOmega is strigtly client side.
[11:57]	Andrew Linden:	which should have been out by then...
[11:57]	Creem Pye:	well, you could try to crash this sim now. it would end the meeting with a bang =P
[11:57]	Johan Laurasia:	yeah, we tried it on two different sims
[11:57]	Johan Laurasia:	one was 1.22.2.89099
[11:57]	Andrew Linden:	Well, that is a good report -- sounds easy to reproduce.
[11:57]	Johan Laurasia:	lol
[11:57]	Andrew Linden:	So it shouldn't be too hard to test and then fix.
[11:57]	Teravus Ousley:	:D
[11:57]	Johan Laurasia:	I'll let Andrew or Simon have the honors
[11:57]	Teravus Ousley:	yay for good repros.
[11:57]	Johan Laurasia:	just rez it, unlink it, then click edit linked parts
[11:57]	Andrew Linden:	Ok good to know Johan.
[11:57]	Rex Cronon:	r u going to crash it right now?
[11:58]	Johan Laurasia:	takes a few seconds.. maybe 5 or so
[11:58]	Arawn Spitteler:	is Crash Me up and running?
[11:58]	Johan Laurasia:	I think you have to touch each prim to get the rotating too
[11:58]	Aeron Kohime:	Crash Me is up last I saw
[11:58]	Aeron Kohime:	I came from ther eto here
[11:58]	Johan Laurasia:	I think it's possibly related to the fact that it's rotating with llTargetOmega()
[11:59]	Simon Linden:	Interesting ... I just tried that bug on a development sim and wasn't able to make it crash
[11:59]	Arawn Spitteler:	Might it have to be a busy sim?
[11:59]	Johan Laurasia:	nope
[11:59]	Johan Laurasia:	when I tested it, I went to an empty mainland sim
[11:59]	Andrew Linden:	If the crash is what I think it is... then it may require that some of the unlinked parts get deleted before the old root prim.
[11:59]	Johan Laurasia:	didnt want to grief anyone
[12:00]	Johan Laurasia:	Simon, did you start the individual prims rotating by touching them first?
[12:00]	Andrew Linden:	So if they are getting deleted in random order then you may have a 1/3 chance of a crash for each original linked set.
[12:00]	Andrew Linden:	er... maybe 2/3... not sure
[12:00]	Simon Linden:	Ah, there's the crash...
[12:00]	Johan Laurasia:	no deleting necessary
[12:00]	Johan Laurasia:	tada!
[12:00]	Johan Laurasia:	thanks ot a week old newbie :)
[12:01]	Andrew Linden:	you going to track that one down Simon?
[12:01]	Teravus Ousley:	haha, my 'accidents' were discovered when I was a n00b also :D
[12:01]	Johan Laurasia:	:)
[12:02]	Simon Linden:	I'll talk with you afterwards Andrew - it's off in a part of the sim that's unmapped by me, just an area with a notation "there be straynge code here"
[12:02]	Andrew Linden:	We really need to extend the Love Machine to include SL Residents.
[12:02]	Johan Laurasia:	'love machine' ?'
[12:02]	Johan Laurasia:	dont have enough of those in SL .. lol
[12:02]	Gaius Goodliffe:	lol
[12:02]	Rex Cronon:	uses batteries?
[12:03]	Andrew Linden:	The "Love Machine" is a system in place in LL that allows employees to show appreciation for others.
[12:03]	Johan Laurasia:	ah
[12:03]	Aeron Kohime:	sound interesting
[12:03]	Rex Cronon:	i thought u were talking about toys. lol
[12:03]	Andrew Linden:	A streamlined method for saying "Thank You" or "I appreciate what you did"
[12:03]	Gaius Goodliffe:	This sounds terribly "San Francisco" to me... :)
[12:03]	Aeron Kohime:	kinda reminds me of the SL score system that was dropped
[12:03]	Johan Laurasia:	Haight is Love.. lol
[12:04]	Johan Laurasia:	oh yeah, ratings
[12:04]	Gaius Goodliffe:	Hehe
[12:04]	Johan Laurasia:	forgot those
[12:04]	Teravus Ousley:	Ratings were a neat market sinc
[12:04]	Andrew Linden:	No, not ratings. Each unit of "love" is only positive.
[12:04]	Teravus Ousley:	:D
[12:04]	Gaius Goodliffe:	There is no bad love. ;)
[12:04]	Johan Laurasia:	why did they get rid of ratings anyways?
[12:05]	Andrew Linden:	Ratings were problematic... people gamed them.
[12:05]	Arawn Spitteler:	Ratings could give bonuses to free accounts
[12:05]	Johan Laurasia:	yeah, what's not gamed...
[12:05]	Gaius Goodliffe:	They were mostly unused by the time 2006 rolled around...
[12:06]	Teravus Ousley:	I got my love rating: In the client about box of 1.19.1(4), "Thank you to the following residents for helping to ensure that this is the best version yet: .... Tee Cramer, Teravus Ousley, Thomas Shikami"
[12:06]	Teravus Ousley:	:D
[12:06]	Johan Laurasia:	woot
[12:06]	Aeron Kohime:	Well done
[12:07]	Andrew Linden:	If the Love Machine were converted to include SL Residents, or if there was some SL equivalent of the Love Machine, then I could enter one item such as "Andrew Linden says thanks to Hughness Osterham for reporting MISC-1261"
[12:07]	Teravus Ousley:	Thanks, I think it was a direct relfection of who submitted a Jira that got looked at
[12:07]	Teravus Ousley:	'reflection'
[12:07]	Andrew Linden:	it would go on record somewhere... an maybe Hughness would recieve some small L$ reward for it.
[12:07]	Gaius Goodliffe:	You could offer L$ bug bounties. But it might drown the JIRA...
[12:07]	Johan Laurasia:	that sounds like a cool idea
[12:08]	Simon Linden:	I gotta run folks - thanks all for coming with the good bug reports and ideas
[12:08]	Teravus Ousley:	haha, L$ patch rewards.
[12:08]	Andrew Linden:	Yeah Gaius, I'm not sure a required bounty would work
[12:08]	Johan Laurasia:	tc Simon, nice meeting ya
[12:08]	Teravus Ousley:	Take care. :D
[12:08]	Simon Linden:	bye
[12:08]	Creem Pye:	yeah especially if 2 people independently reported a bug, but the first person wrote a poor summary/repro
[12:08]	Rex Cronon:	bye simon
[12:08]	Andrew Linden:	but it would be nice if we could easily distribute small bounties for small help
[12:08]	Aeron Kohime:	bye Simon
[12:08]	Gaius Goodliffe nods.
[12:08]	Teravus Ousley:	I should probably go also. Getting late.
[12:09]	Gaius Goodliffe:	Yeah, it would have to be judged intelligently. Which will ensure people will scream "favoritism"
[12:09]	Rex Cronon:	bye everybody
[12:09]	Andrew Linden:	anyway, many people like the appreciation of a unit of "love" more than any monetary value it might have.
[12:09]	Rex Cronon:	have fun
[12:09]	Aeron Kohime:	I would help patch but they would most likely break things more then they fixed.
[12:09]	Johan Laurasia:	Great work on openSim Teravus... I've been tinkering around with it lately
[12:09]	Teravus Ousley:	Thanks again for holding this meeting Andrew :D
[12:09]	Arawn Spitteler:	Acknowledgement, of which Money is only a Fungible Unit
[12:09]	Andrew Linden:	Ah yes, time to go.
[12:09]	Andrew Linden:	Thanks for comming everyone.
[12:09]	Creem Pye:	thanks for your time
[12:09]	Johan Laurasia:	Thank Andrew :)
[12:10]	Gaius Goodliffe:	Thank you sir. :)
[12:10]	Teravus Ousley:	ty, johan :D
[12:10]	Aeron Kohime:	Thankyou Andrew
[12:10]	Aeron Kohime:	Thanks for the meeting everyone
[12:10]	Arawn Spitteler:	We need a DOC project, for Jira
[12:10]	Gaius Goodliffe:	And THIS is why we need lllUnSitTarget. :)
[12:10]	Johan Laurasia:	heh
[12:10]	Teravus Ousley:	haha
[12:10]	Gaius Goodliffe gets off the table.
[12:10]	Johan Laurasia:	hmm, llUnsitTarget() sounds useful
[12:11]	Arawn Spitteler:	Standing up changed, at some point
[12:11]	Johan Laurasia:	I'm tired of getting launched sometimes when I stand up
[12:11]	Andrew Linden changed standing up in Havok4.