User:Andrew Linden/Office Hours/2008 11 18

Transcript of Andrew Linden's office hours:

[10:57]	Object: Hello, Avatar!
[10:58]	Kitto Flora:	Is there a new meet place?
[10:58]	Andrew Linden:	yes
[10:58]	Andrew Linden:	I'm about to leave a reminder object with floating text
[10:58]	Andrew Linden:	Denby... somewhere
[10:58]	Andrew Linden:	213/45
[10:58]	Kitto Flora wtches
[10:59]	Kitto Flora:	Simon's place?
[10:59]	Andrew Linden:	yeah
[10:59]	Kitto Flora:	Close
[11:00]	Kitto Flora:	No LM giver?
[11:00]	Andrew Linden:	hehe, no 'cause I'm lame
[11:00]	Andrew Linden:	actually, I've never written one
[11:00]	Cummere Mayo:	written what?
[11:00]	Kitto Flora gave you New Havok4 meet place.
[11:00]	Cummere Mayo:	and hellow andrew. been a while :-/
[11:01]	Kitto Flora gave you GiveAllContents 2.1, set to give to ANYONE.
[11:02]	Cummere Mayo:	we being moved just for the day or for good?
[11:02]	Andrew Linden:	ok, how's that?
[11:02]	Andrew Linden:	does it hand out a landmark?
[11:02]	Gaius Goodliffe:	yup
[11:02]	Cummere Mayo:	no it handed out a box
[11:03]	Gaius Goodliffe:	eerr oops
[11:03]	Gaius Goodliffe:	hehehe
[11:03]	Gaius Goodliffe:	At least it hands out something XD
[11:03]	Andrew Linden:	gah
[11:03]	Kitto Flora:	Its giving a folder with the script in it
[11:03]	Cummere Mayo:	hmm my other avi has the lm giver script
[11:04]	Rex Cronon:	hello
[11:04]	Kitto Flora:	Set the candmarrk Copy
[11:04]	Kitto Flora:	landmark
[11:04]	Gaius Goodliffe:	Just check if inventory name != llGetScriptName
[11:04]	Cummere Mayo:	have you updated the calandar and wiki andrew?
[11:04]	Kitto Flora:	Yay!
[11:04]	Andrew Linden:	ok I fixed the script
[11:05]	Kitto Flora:	Works
[11:05]	Gaius Goodliffe:	And there was much rejoicing...
[11:05]	Andrew Linden:	Ok, I'm off to the new location
[11:06]	Simon Linden:	Hi Rex, glad you made it
[11:06]	Kitto Flora:	Guess the furniture truck didnt make it yet
[11:06]	Rex Cronon:	hi simon. hi everybody
[11:06]	Simon Linden:	yeah, I just got one of these simple expando-tables from OmniCorp
[11:06]	Rex Cronon:	btw i don't see anybody
[11:07]	Andrew Linden:	I don't know why people make houses in SL with roofs.
[11:07]	Cummere Mayo:	um privacy?
[11:07]	Simon Linden:	That was interesting
[11:07]	Rex Cronon:	they r afraid that flying birds might leave a present:)
[11:07]	Andrew Linden:	oof
[11:07]	Gaius Goodliffe:	hehe
[11:07]	Gaius Goodliffe:	need traffic lights there
[11:07]	Andrew Linden:	maybe I'll stand
[11:07]	Cummere Mayo:	i like having a roof to help discourage the um.... voyers
[11:07]	Gaius Goodliffe:	Shoudl be safe now XD
[11:07]	Andrew Linden:	simon, is this an openspace region?
[11:08]	Kitto Flora:	Weird - tables running slow but LSL isnt too bad here
[11:08]	Simon Linden:	I'm not sure, lemme check...
[11:08]	Cummere Mayo:	if it is.... then its in violation of LL's own "policy"
[11:08]	Simon Linden:	Nope, we're safe from the Policy Police. It's a standard sim
[11:09]	Gaius Goodliffe:	9700 objects... definitely not an openspace
[11:09]	Andrew Linden:	alright
[11:09]	Andrew Linden:	ok, so welcome everyone to the new office hours
[11:09]	Andrew Linden:	Thanks to Simon for providing the location
[11:09]	Rex Cronon:	hiii
[11:10]	Andrew Linden:	My name is Andew Linden and I'm here to answer any questions you may have.
[11:10]	Kitto Flora:	Any news?
[11:10]	Andrew Linden:	Server-1.25 has hit some snags
[11:10]	Andrew Linden:	and is not yet released
[11:10]	Andrew Linden:	lemme check the daily update email...
[11:11]	Andrew Linden:	it appears that there are still two problems blocking server-1.25
[11:12]	Andrew Linden:	one of them is what knocked out logins last week... Thursday or Wednesday
[11:12]	Arawn Spitteler:	The server knocked out logins?
[11:12]	Andrew Linden:	That's two problems out of about 7. Most of them have been fixed already.
[11:13]	Andrew Linden:	No, not the simulator, but some changes to some other servers that must also proceed with server-1.25
[11:13]	Kitto Flora:	Insufficient testing in Preview?
[11:13]	Simon Linden:	Mostly problems with our test grids not being good substitutes for the big grid
[11:13]	Andrew Linden:	in particular the server we call "backbone" which is really an python app that serves simple XML info via HTTP
[11:13]	Arawn Spitteler:	I understand the porblem is that development environments can't alpha for loading
[11:13]	Gaius Goodliffe:	Teen Grid for testing? ;)
[11:14]	Andrew Linden:	it appears that backbone has a memory leak that was not bad enough on the test grids to be visible
[11:14]	Kitto Flora:	The Aditi test grid is not condusive to content maker testing
[11:15]	Arawn Spitteler imagines a test sim in Teen Grid named "Don't Hurt Me."
[11:15]	Gaius Goodliffe:	XD
[11:15]	Andrew Linden:	backbone is built with co-routines (using eventlet which uses greenlet, which are python libs for co-routine stuff)
[11:15]	Kitto Flora:	Nor does it entice general population for anything other than designing clothing
[11:15]	Gaius Goodliffe:	Hey, if we can get out kids to wash dishes, take out trash, and fetch remotes, we ought to be able to use them to test software. :)
[11:16]	Gaius Goodliffe:	*our kids
[11:16]	Simon Linden:	I can't get mine to wash dishes or take out trash
[11:16]	Gaius Goodliffe:	XD
[11:16]	Arawn Spitteler:	That's what wives are for, getting someone to wash dishes and take out trash
[11:16]	Gaius Goodliffe:	:o
[11:17]	Andrew Linden:	Looks like there are some people woking on both server-1.25 blocker problems.
[11:17]	Kitto Flora:	Meanwhile - do you know anything about height-plane corruption due to sim version changes?
[11:18]	Andrew Linden:	Arawn, looks like SVC-2931 has leaked into our internal jira via a different bug. Looks like i'm going to have to write up an opinion on that issue soon. Hopefully I'll get to it today.
[11:18]	Arawn Spitteler:	Height Plane Corruption? Which server version.
[11:18]	Andrew Linden:	Heightfield corruption? No I haven't heard of it yet.
[11:18]	Andrew Linden:	Hrm...
[11:19]	Kitto Flora:	Well ,whatever the server version has been the last week or so
[11:19]	Arawn Spitteler:	I think it's SVC-3408; that someone's clamped down the distance further.
[11:19]	Arawn Spitteler:	I fuigure the traditional link distance for an avatar is 300 meters to an axis.
[11:19]	Cummere Mayo:	server version has been the same the last week for most regions that it was before?
[11:20]	Andrew Linden:	There was a small deployment of server-1.25 last week.
[11:20]	Andrew Linden:	Dunno how long it lasted.
[11:20]	Simon Linden:	There were a few machines pushed to 1.25 - one rack, I think
[11:20]	Kitto Flora:	What we found is that previously flattened land suddenly has 'waves' in it - undulations of up to 0.5M height and 10's of meters period
[11:20]	Andrew Linden:	I figure all the server-1.25 regions have been redeployed to 1.24
[11:20]	Simon Linden:	I believe they were reverted, however, after the back-end problems were seen
[11:20]	Arawn Spitteler:	I think Darling was on the phone to support, when it got pulled
[11:20]	Andrew Linden:	What regions?
[11:21]	Kitto Flora:	(This is in 1.24 - afaik problem predated 1.25 deploy)
[11:21]	Kitto Flora:	Region affected is Wish
[11:21]	Andrew Linden:	Approximately when did the corruption happen?
[11:21]	Kitto Flora:	Noticed about a week ago
[11:22]	Andrew Linden:	ack... anything older than a few days is hard to track down
[11:22]	Arawn Spitteler:	What corruption is seen?
[11:23]	Kitto Flora:	What we found is that previously flattened land suddenly has 'waves' in it - undulations of up to 0.5M height and 10's of meters period
[11:23]	Cummere Mayo:	arawn read up to time stamp 1120]
[11:23]	Andrew Linden:	looks like Wish was moved to a new process ID on the 16th
[11:23]	Gaius Goodliffe:	Did the height show differently in llGround?
[11:23]	Andrew Linden:	and th 12th
[11:23]	Andrew Linden:	and the 9th (or 10th)
[11:23]	Kitto Flora:	Cant tell - no previous records of llGorund()
[11:24]	Kitto Flora:	Thing is - anyone else reporting such problems?
[11:24]	Andrew Linden:	Kitto, how long was the terrain corruption present?
[11:24]	Andrew Linden:	That is, how long until someone fixed it?
[11:24]	Kitto Flora:	Several days between when the owner complained at me
[11:25]	Kitto Flora:	to when I finally figuted out what and where the complaint was about
[11:25]	Andrew Linden:	ah ok that is good, then I might still have some snapshots of the bad terrain
[11:25]	Andrew Linden:	do you happen to know any of the dates within that period?
[11:25]	Kitto Flora:	I then used large size smooth brush to flatten the land again
[11:26]	Kitto Flora:	And theres another area that has bumps in it now - Owner has to look see if they are 'new bumps'
[11:26]	Andrew Linden:	perhaps someone dropped a notecard on you and it has a creation date timestamp?
[11:26]	Kitto Flora looking...
[11:27]	Kitto Flora:	No notecards
[11:27]	Cummere Mayo:	this might sound like a crazy question but what viwer version were the reoprts using?
[11:27]	Andrew Linden:	we're up on the roof Chaley
[11:27]	Kitto Flora:	IM log only goes back a day
[11:28]	Kitto Flora:	Using Second Life 1.21.6 (99587) Oct 14 2008 17:42:25 (Second Life Release) viewer
[11:28]	Andrew Linden:	ok, anybody have any other topics to raise?
[11:28]	Rex Cronon:	i have a question
[11:29]	Kitto Flora:	Durn - no email either
[11:29]	Arawn Spitteler:	llTPAgent: I don't see much potential for Griefing, there.
[11:29]	Rex Cronon:	has it ever happened to anybody to be in a sim, and when u r on the ground u start sliding to the East?
[11:29]	Gaius Goodliffe:	Oh yeah skating XD
[11:30]	Andrew Linden:	hrm... that sounds like a problem I've heard about
[11:30]	Arawn Spitteler:	I was sliding a little, just a little ago, but no special direction
[11:30]	Rex Cronon:	on a flat plane
[11:30]	Gaius Goodliffe:	I assume it's just windy. ;)
[11:30]	Arawn Spitteler:	I think a square meter has a sweet spot, that we gravitate to, but it changes when we turn
[11:30]	Kitto Flora:	Sliding downhill, maybe .
[11:30]	Andrew Linden:	I think it may be caused by NaN (not a number) injection into the physics engine by a scripted object
[11:31]	Kitto Flora:	Sliding while turning too, perhaps
[11:31]	Arawn Spitteler:	I wasn't wearing any physics scripts that I know of.
[11:31]	Rex Cronon:	from what i heard it is viral
[11:32]	Andrew Linden:	right, the NaN infection is viral, and the physics engine deals with it by clamping the object velocity to something in the positive-X direction (hence East0
[11:32]	Andrew Linden:	there is at least one known injection method. I need to fix that bug.
[11:32]	Andrew Linden:	that is, known to me
[11:32]	Rex Cronon:	ok. i wasn't sure if i was being made fun of, regarding the viral part
[11:33]	Arawn Spitteler:	I don't know how a virus could infect from a script
[11:33]	Andrew Linden:	I think any NaN infected object also introduces NaN's to any obejct it touches.
[11:33]	Gaius Goodliffe:	Monkey-wrenching the physics engine?
[11:33]	Andrew Linden:	1 + NaN = NaN, 1 * NaN = NaN etc
[11:34]	Simon Linden:	It's basically a bad number that propigates into other objects when the value is used in computations
[11:34]	Kitto Flora:	Inapropriate data being imported in a data set?
[11:34]	Simon Linden:	Object A has the NaN, it hits object B, and we use the value from A to figure momentum or something, and thus B ends up with bad value
[11:35]	Andrew Linden:	no, the NaN is introduced by providing some very large (or small) numbers into some calculation that the simulator does
[11:35]	Andrew Linden:	such that the product is NaN
[11:35]	Gaius Goodliffe wonders how A gets a NaN to begin with, and whether this can be used in making an infinite improbability drive.
[11:35]	Andrew Linden:	so we need to find that calculation spot, and do a sanity check on the input
[11:35]	Arawn Spitteler:	Sounds like a calculation is just coming up with things too small to calculat4e
[11:36]	Gaius Goodliffe:	Divisiion by effectively zero?
[11:36]	Andrew Linden:	you used to be able to generate a NaN via llMoveToTarget()... supplying sufficiently bad numerical values
[11:36]	Andrew Linden:	I think this one is through a different vector.
[11:37]	Andrew Linden:	actually, I think division by zero causes a crash
[11:37]	Kitto Flora:	Maybe a few sanity check points scattered around to find the cause?
[11:38]	Arawn Spitteler:	No Exception Handlers?
[11:38]	Simon Linden:	The cases I saw before were large numbers that got multiplied
[11:38]	Gaius Goodliffe:	(One hopes that by "crash" one means "exception thrown")
[11:38]	Arawn Spitteler:	"If you need more than this number, why?"
[11:39]	Andrew Linden:	that is, I think division by zero hits an assert() that just causes the process to crash
[11:39]	Kitto Flora:	Another general question - does 'Advanced/Consoles/Fast timers show data for whats happening in the clinet?
[11:40]	Kitto Flora:	client
[11:40]	Kitto Flora:	(or the sim?)
[11:40]	Andrew Linden:	yes Kitto
[11:40]	Andrew Linden:	the client
[11:40]	Andrew Linden:	I haven't really played with those
[11:40]	Kitto Flora:	OK. Would any of that data be affected by changes between LSL2 and MONO scripts?
[11:40]	Andrew Linden:	hrm... I'm trying to remember the Linden who wrote those...
[11:41]	Arawn Spitteler:	llTeleportAgent, is that possible to use in a Griefing Mode, beyond what happens with llTeleportAgentHome?
[11:41]	Andrew Linden:	No, LSL2 and Mono should not affect the fast timer stats
[11:42]	Kitto Flora:	Arawan - I'd think about the same - cept you could zap the agent to some other weird place
[11:42]	Andrew Linden:	Runitai Linden wrote th fast timer stuff, I think.
[11:43]	Arawn Spitteler:	Teleport Agent would be useful, if your build is a maze of weird places, but it's held up, and I think it's for fearof Greifing
[11:43]	Kitto Flora:	Andrew: I have a customer who swears they get 'vast improvement' having converted all scripts in a sim to MONO, and chalims that Fast Timers are the evidence for the 'vast improvemnet'. To me this seems like 'theres no connection'. Does it make any sense to you?
[11:43]	Andrew Linden:	Yes, I think there were griefing potentials in it. I still need to talk to someone who remembers what the issues were.
[11:43]	Gaius Goodliffe:	I would think it would be okay as long as it has the same kind of restrictions as llUnSit or the like regarding land owner and such
[11:44]	Andrew Linden:	Kitto, the fat timers have info about the render pipeline.
[11:44]	Morgaine Dinova:	'Evening :-)
[11:44]	Andrew Linden:	So no, it does not make sense.
[11:44]	Arawn Spitteler:	It's a matter of policing the parcel owners, with Payment Info on file, rather than us Sandbox Users
[11:44]	Kitto Flora:	OK.
[11:44]	Rex Cronon:	hi
[11:44]	Arawn Spitteler:	Hio, Morgaine
[11:44]	Andrew Linden:	hello Morgaine
[11:45]	Kitto Flora:	I suspect its more like they also updtaed client of sim at around the same time and that made a diff to fast timers.
[11:45]	Andrew Linden:	Ok, so how could llTeleportAgent() be used to grief?
[11:45]	Kitto Flora:	So far aat best I've seen 10% improvement when converting to MONO. And thats 'down in the noise'.
[11:46]	Andrew Linden:	That is, a script feature that could trigger an avatar teleportation event.
[11:46]	Arawn Spitteler:	Some guy's claiming that SVC-2931 and SVC-3408 are not problems, because we should have other functions for that, but I don't think TPAgent would be as eficient or convenient, for a lot of apps.
[11:46]	Cummere Mayo:	if you can somehow get that to target another avi you could tp them whever the hell you wnated
[11:46]	Andrew Linden:	Naturally, you'd need to give permissions to be teleported, right?
[11:46]	Gaius Goodliffe:	With appropriate restrictions ala llUnSit, I don't think it really can. Without that, it's problematic.
[11:46]	Simon Linden:	Mono should speed up general execution and calculations, but the real script bottlenecks are rezzing and moving objects. That's using the same code in LSL or Mono
[11:46]	Morgaine Dinova:	An improvement of 10% doesn't tally with Cory's old "400 times improvement" claim
[11:47]	Andrew Linden:	the 400% improvement was misleading, yes. "Up to 400%..."
[11:47]	Gaius Goodliffe:	All benchmarks are misleading.
[11:47]	Andrew Linden:	certain types of code will run much much faster on Mono
[11:47]	Morgaine Dinova:	No no, he didn't say 400%, he said 400 times
[11:48]	Morgaine Dinova:	40,000% ;-)
[11:48]	Andrew Linden:	in particular, math-heavy code that doesn't call many llFoo() functions
[11:48]	Gaius Goodliffe nods.
[11:48]	Kitto Flora:	the 400 times was for math calcs. Vehicles only use a few of those. Dominated by Events
[11:48]	Andrew Linden:	oh right... 400X
[11:48]	Morgaine Dinova:	Aye, not surprised.
[11:48]	Gaius Goodliffe:	I've seen code what was at least 20 times faster, doing a whole bunch of munching on data.
[11:48]	Arawn Spitteler:	TPAgent can always be countered with <shift><ctrl>h, if I've spelled it right.
[11:48]	Andrew Linden:	LSL vehicle calls are really just hooks for setting params in a "physcis action".
[11:49]	Morgaine Dinova:	But in that case, are we saying that all the effort of switching to Mono has provided only small-scale improvements?
[11:49]	Andrew Linden:	When using vehicles you are primarily leaning on the physics engine,
[11:49]	Andrew Linden:	except for any event handling you're doing in the script.
[11:49]	Cummere Mayo:	depends on the types of scrips morgaine, but yes i think thats the case overall, BUT those small improvements DO add up
[11:49]	Gaius Goodliffe:	It's most definitely not only provided small-scale imprevmenets!
[11:50]	Simon Linden:	Mono is also much, much more stable -- using it should remove a lot of crash modes
[11:50]	Gaius Goodliffe:	Some scripts might only see small-scale improvements. Others are HUGELY improved.
[11:50]	Andrew Linden:	Morgaine, yes I think there were some small improvements in average in script performance
[11:50]	Andrew Linden:	right, some see large improvements
[11:50]	Morgaine Dinova:	Ah, good.
[11:50]	Andrew Linden:	and there was a lot of internal overhaul involved that is supposed to let us move forward in new and interesting ways
[11:51]	Cummere Mayo:	i lvoe the qualifiers you add in andrew
[11:51]	Andrew Linden:	that said, I don't think Babbage Linden's team is working on those "new and interesting" things yet
[11:51]	Arawn Spitteler:	I recall somethig about the sim requiring only one copy of each script
[11:51]	Kitto Flora:	Improved llSensor() and link messages is what I need :)
[11:51]	Gaius Goodliffe:	Trying to talk about how things are overall is guarenteed to generate misleading info. All generalizations are false. ;)
[11:51]	Andrew Linden:	Cummer, I have a little flaw... I try to say things that are "correct"
[11:51]	Andrew Linden:	so I have to add a lot of qualifiers
[11:51]	Cummere Mayo giggles
[11:51]	Gaius Goodliffe nods.
[11:52]	Cummere Mayo:	its ok. I like hoensty.
[11:52]	Gaius Goodliffe:	No true statement has less than three clauses and 6 conditionals. ;)
[11:52]	Andrew Linden:	Yes Arawn is right, there were some improvements on the average memory footprint for scripts in the MONO project
[11:52]	Arawn Spitteler:	This statement is true, unless otherwise.
[11:52]	Andrew Linden:	which should contribute to general speedup
[11:52]	Cummere Mayo:	you see allot of it in the "rank and file" lindens and not so much in certain other ones
[11:52]	Morgaine Dinova:	I take it that arrays will finally make it into LSL, given the Mono target. :-) Don Linden told us that they were "working already, just tidying up now" ... back in 2005 ;-))))
[11:53]	Gaius Goodliffe:	Hmmm.
[11:53]	Andrew Linden:	ooof... that 2005 statement never came to fruition
[11:53]	Kitto Flora:	Andrew, The Heightfield bumps corruption was first noticed around November 1st. But when it really happened could have been weeks before that.
[11:53]	Gaius Goodliffe:	I think it would make more sense to get C# and Python working than adding stuff like that to LSL.
[11:53]	Andrew Linden:	Don Linden knows more. I know very little about it.
[11:54]	Andrew Linden:	Yes I think the C# is the next big project that Babbage wants to work on, but not positive.
[11:54]	Morgaine Dinova:	Gaius: I agree, but publicly at least, that';s not happening any time soon :-(
[11:54]	Gaius Goodliffe:	Soon is such a relative term... ;)
[11:55]	Simon Linden:	Yes, we're proably not going to add language features, but functions that interact with SL are possible
[11:55]	Morgaine Dinova chuckles
[11:55]	Morgaine Dinova:	Well 2005 was recent ;-)
[11:55]	Andrew Linden:	Simon, could you elaborate on that?
[11:55]	Gaius Goodliffe:	Geologically speaking... :)
[11:55]	Morgaine Dinova:	Hehe
[11:55]	Arawn Spitteler:	Most of the oldbies I meet started in 2006
[11:56]	Simon Linden:	I mean that we're more likely to add a llDoSomethingInteresting() function than adding a better array type to LSL
[11:56]	Gaius Goodliffe:	Right. Build the library...
[11:56]	Gaius Goodliffe:	That will benefit any language we may use in the future.
[11:57]	Gaius Goodliffe:	Might I suggestion llUnSitTarget *plug plug plug*
[11:57]	Andrew Linden:	The llC# or llPython effort would be to provide a new interface to the same LSL features.
[11:57]	Cummere Mayo proposes a		SmackLLExecInTheHeadWithFryingPan() function
[11:57]	Andrew Linden:	There were some internal discussions going on about it several months ago, closer to the impending Mono release.
[11:58]	Kitto Flora:	llUnSitTarget or llTPAgent wuld have been good for the getting people out of vehicles. But I found another way around the trapping problem
[11:58]	Arawn Spitteler:	llUnSitTarget? llUnSit is already given a target.
[11:58]	Andrew Linden:	At least one Linden argued that LSL has a very terse interface API, and that it would take more code to do the same thing in Python, or C#
[11:58]	Kitto Flora agrees with one Linden
[11:59]	Morgaine Dinova:	Well from the AWG angle, sim scalability is a big issue, and LSL's lists are probably soaking up an inordinate amount of CPU time since they're O(N) instead of O(1), and the algorithms used by users to make that even worse than O(N). So if profiling points the finger at list operations, it might be an idea to add arrays to decrease sim loading a bit.
[11:59]	Andrew Linden:	however, if a suitable python or C# lib were created then we could maintain the terseness of typical LSL scripts.
[11:59]	Gaius Goodliffe:	Arawn, no, not really... not as in a position ala llSitTarget
[11:59]	Andrew Linden:	I'd like to be able to try to optimize the exsiting script engine. I suspect there are some easy things to do in there
[12:00]	Arawn Spitteler:	I know I'd like a DetectedTarget
[12:00]	Andrew Linden:	however others say that we've probably already found the easy bits.
[12:00]	Simon Linden:	On an aside, I did add some internal metrics code last week that will make it easier for us to figure out what LSL functions take up too much time. The initial results were as expected - llRezObject() and llSetPos(). This was on a sim that sat empty at 20 fps
[12:00]	Morgaine Dinova:	That's interesting Simon
[12:00]	Simon Linden:	... some rezzer objects were the bottleneck
[12:01]	Kitto Flora:	Simon can you put them in the Client Stats panel?
[12:01]	Morgaine Dinova:	Sim stats you mean :-)
[12:01]	Simon Linden:	It's not there, but we did discuss how it would be nice to expose that information
[12:01]	Arawn Spitteler:	RezRElative is another idea. Don't I recall talk of using primal objects, for the sims to preload?
[12:01]	Rex Cronon:	have u tried rezzing both scripted and unscripted objects
[12:02]	Kitto Flora:	Sim stats display in the client
[12:02]	Simon Linden:	I think 1.25 is going to have one feature where the 'top scripts' will include attachments, so an AV will show up in the list
[12:02]	Rex Cronon:	to me it seems that rezzing scriptted objects is a lot slower
[12:02]	Kitto Flora:	Rather than a tool running on the server
[12:02]	Gaius Goodliffe:	How big of a difference does temp-on-rez actually make?
[12:02]	Morgaine Dinova:	Simon: oh, that's so cruel :-))))))
[12:02]	Gaius Goodliffe:	:D
[12:02]	Gaius Goodliffe:	/yay
[12:02]	Simon Linden:	None at all -it's our parsing time and the overhead of adding it to the physics engine that hurts
[12:02]	Gaius Goodliffe has wanted avatar stats for ages.
[12:03]	Kitto Flora:	Theres a lot of pressure on trying to debug sim loads at present, but its driven by OS sims. And will eventually vanish?
[12:03]	Gaius Goodliffe:	Ag
[12:03]	Gaius Goodliffe:	ah
[12:03]	Simon Linden:	The OS sims show the problem the most, but the work on them will benefit standard regions too
[12:03]	Kitto Flora:	Yes
[12:03]	Cummere Mayo:	*Sighs* rate things are going within 2 months many os sims will be gone
[12:03]	Andrew Linden:	yeah, temp-on-rez is just a bit stored in the object. We reap it after some time if that bit is set. Otherwise it is like any other object.
[12:04]	Gaius Goodliffe nods.
[12:04]	Kitto Flora:	Script count limiting - by parcel and Av - would be great
[12:04]	Kitto Flora:	Like Prim limits
[12:05]	Gaius Goodliffe:	How about rezzing phantom objects? Those won't load the physics engine as much, right?
[12:05]	Simon Linden:	I have some ideas on speeding up rezzers ... perhaps we can cache the parsing results since they tend to make the same thing over and over
[12:05]	Arawn Spitteler:	Preload Object?
[12:05]	Gaius Goodliffe nods.
[12:05]	Gaius Goodliffe:	A boon for guns.
[12:05]	Simon Linden:	preloading would be interesting
[12:05]	Andrew Linden:	That reminds me about some project i was going to work on to make limits on scripts and other object-content assets easier to implement...
[12:05]	Arawn Spitteler:	In making my polymorphic Teleporter, I thought of another use for that.
[12:05]	Andrew Linden:	one of these days I'll get to i.
[12:06]	Rex Cronon:	what would be great. it would be if a script can interogate an object and obtain a list of all the scripts is using:)
[12:06]	Andrew Linden:	alright, looks like this hour is up. I've got to head out.
[12:06]	Andrew Linden:	Thank you all for coming.
[12:06]	Cummere Mayo:	take care andrew
[12:06]	Gaius Goodliffe:	Thank you, Andrew.
[12:06]	Simon Linden:	Me too ... thanks everyone for coming
[12:06]	Gaius Goodliffe:	Thank you, Simon.
[12:06]	Kitto Flora:	Andrew - according to Linden Blog - script count limits are supposed to be on their way :)
[12:06]	Morgaine Dinova:	Oh a more long-term level, is there any move towards splitting script execution off into a separate Script Domain (just like a lot of stuff is being removed from sim into AD), with server pools etc? Again, scalability being the driving force, get load off sim.
[12:06]	Cummere Mayo:	simon do you have some time?
[12:07]	Morgaine Dinova:	Cheers all
[12:07]	Gaius Goodliffe:	I know they added a limit earlier this year on the number of scripts an object can contain...
[12:07]	Kitto Flora:	Lunchtime!
[12:07]	Simon Linden:	Cummere - a few minutes
[12:07]	Arawn Spitteler:	Scripts off sim mean communications load, but is otherwise attractive.