User:Andrew Linden/Office Hours/2010 08 20

Transcript

[16:03]	Moon Metty:	hi Simon :)
[16:04]	Cummere Mayo:	hey simon :)
[16:04]	DogWomble Dollinger:	hello simon
[16:04]	Simon Linden:	hello
[16:04]	Cummere Mayo:	so simon have you seen the latest emerald scandel?
[16:04]	Techwolf Lupindo:	/me smiels
[16:04]	Stryker Jenkins:	I see a blue dot on my minimap... could it be?
[16:04]	Simon Linden:	no, what's up?
[16:04]	Stryker Jenkins:	haha
[16:04]	reddot99 Republic:	hey andrew
[16:04]	Cummere Mayo:	let me grab the link fast
[16:04]	DogWomble Dollinger:	hi andrew
[16:04]	Romaq Rosher:	Howdy!
[16:04]	Andrew Linden:	hello
[16:05]	Andrew Linden:	Simon and I were in the middle of a code review -- we almost forgot to show up
[16:05]	reddot99 Republic:	lol
[16:05]	Moon Metty:	hi hi :)
[16:05]	Cummere Mayo:	http://blogs.secondlife.com/thread/34098
[16:06]	Stryker Jenkins:	how is live in the land of CODE...
[16:06]	Morgaine Dinova:	Hi Andre/Simon :-)
[16:06]	Andrew Linden:	I had a noteworthy experience this week... a unit test caught an error in my code and I fixed it immediately
[16:07]	Andrew Linden:	that is what unit tests are supposed to do
[16:07]	Moon Metty:	:)
[16:07]	Gramma Fiddlesticks:	Yay Andrew!
[16:07]	Morgaine Dinova:	Groupies is in the middle of a code review too --- we reviewed Emerald's login page and confirmed that they're using Emerald users as a botnet -- http://secondslog.blogspot.com/
[16:07]	Cummere Mayo:	eeks what kind of error ?
[16:07]	Andrew Linden:	so it means we're starting to get enough unit tests written that their benefit is starting to get noticable
[16:07]	Moon Metty:	is that what Vektor has been working on?
[16:07]	reddot99 Republic:	you mean like obvious bugs?
[16:08]	reddot99 Republic:	SNOW-810
[16:08]	DeedtoGroupFail Error	popup missing from xml files
[16:08]	Andrew Linden:	nothing big, I had forgotten to init a class data member in the last constructor
[16:08]	Andrew Linden:	so I got three out of four correct, and the unit test was using the 4th contructor type
[16:09]	Andrew Linden:	lets see... other small news...
[16:09]	Andrew Linden:	I've been merging a bunch of work so I can put it all up for internal testing
[16:09]	reddot99 Republic:	anything on SEC -730?
[16:09]	Andrew Linden:	and I think I'm just about merging stuff
[16:10]	reddot99 Republic:	where sims crash from rez queue overflowing
[16:10]	Andrew Linden:	SEC-730 is a crash mode in havok -- dunno if that is fixed, but I don't think it is in my pile
[16:11]	Andrew Linden:	I'd have to ask falcon
[16:11]	Simon Linden:	1.42 has some better code for dealing with mass rezzing, but I've seen a few crashes on it that are still related to that type of attack
[16:11]	reddot99 Republic:	i think easy solution would be add more ram
[16:11]	Andrew Linden:	most of my stuff is just cleanup, but some of it is theoretical small optimizations
[16:11]	Andrew Linden:	every little bit helps, as they say
[16:12]	reddot99 Republic:	simon, it doesnt have to be an 'attack' on the server for it to occur
[16:12]	Simon Linden:	We may need to add a region-wide rezzing limit .. should be high enough to never be hit, but if a bunch of objects start at the same time, that might catch it
[16:12]	Andrew Linden:	no progress on region crossings (at least not by me)
[16:12]	Andrew Linden:	but I should start helping with that project next week
[16:13]	reddot99 Republic:	well, hope that people dont try to go to nonetemp bullets in the meantime
[16:13]	Stryker Jenkins:	cool...
[16:13]	reddot99 Republic:	otherwise asset server will probably be choking at some point on it
[16:13]	Tegg Bode:	cleanup and small optimisations is good, that's where often little mistakes are found or new concepts born ㋡
[16:13]	Rex Cronon:	hello everybody
[16:14]	DogWomble Dollinger:	hekki rex
[16:14]	DogWomble Dollinger:	*hello
[16:14]	Rex Cronon:	hello dogwomble
[16:14]	Stryker Jenkins:	I was just at Michaels OH... and we talked a bit about having larger prim sizes as standard... say 16x16 in line with land sizes. Do you have an inshight on that?
[16:15]	Andrew Linden:	I don't want to boost default prim size limits myself
[16:15]	Andrew Linden:	not until we solve prim encroachment and collisions across region boundaries
[16:15]	reddot99 Republic:	i thought havok 7 would help with that
[16:16]	Andrew Linden:	however, I'm aware that the pressure for that request might eventually build until there is a consensus to boost the limit
[16:16]	CPU Core:	Andrew give it to islands.. they dont have prim encroachment issues
[16:16]	Liisa Runo:	we used to have collisions over border, but it got removed
[16:16]	Andrew Linden:	and that would increase the priority of fixing those two
[16:16]	Andrew Linden:	but for now the stuff on my plate appears to be:
[16:16]	Andrew Linden:	(1) take a pass at speeding up region crossings
[16:16]	Andrew Linden:	(2) speed up the interest list
[16:17]	reddot99 Republic:	actually, could you make the mesh for megaprims more dense? like more points, because it'd prevent getting inside them so easily i think
[16:17]	Andrew Linden:	content on private islands can be taken to inventory, carried to the mainland, and rezzed
[16:17]	Moon Metty:	cutting the interest list into smaller pieces?
[16:17]	Cummere Mayo:	I heard region crossings are getting a boost by tkaing some proccessing off the sim's cpu's laod and placing it on an outside cpu?
[16:18]	Romaq Rosher:	/me puzzles, "Interest List".
[16:18]	Tegg Bode:	We could doo without noobs rezzing 128m physical spheres everywhere
[16:18]	CPU Core:	people build with megaprims anyway so thats no argument
[16:18]	Morgaine Dinova:	I just heard yesterday that the interest list is 2D only. Ouch. And I thought a skybox up high kept the interest list sparse.
[16:18]	Simon Linden:	some of the processing will be threaded in 1.42
[16:18]	Andrew Linden:	Cummere, that threaded object rez code is done an will probably help
[16:18]	Andrew Linden:	but is not part of the current sprint that is focusing on region crossing optimizations
[16:19]	Andrew Linden:	Morgaine, correct the interestlist code is 2D atm.
[16:19]	Stryker Jenkins:	does speeding up crossing also help with falling though the floor / land when walking or riding?
[16:19]	Moon Metty:	hmmm, i guess vwr-5588 is related to that
[16:19]	Objects within	draw distance not rendered when standing away some meters
[16:19]	Moon Metty:	should that be moved to SVC?
[16:19]	Cummere Mayo:	oh. I woulda thought it would be since part of the speed loss was the processing time for scripts/attachments etc :-/
[16:19]	Morgaine Dinova:	Andrew: is making it 3D on any scrum's Backlog?
[16:20]	reddot99 Republic:	could you make an option to have something like a 'hypercube' as somebody put it, so like on an island, you drive a vehicle off a border and you can it warp to other side instead of offworld?
[16:20]	Andrew Linden:	Stryker, I think you're seeing the "objects don't collide across region boundaries" problem (when falling through the floor on region crossing)
[16:20]	Stryker Jenkins:	right...
[16:20]	Simon Linden:	Morgaine - I don't think that's specifically called out yet, but it's on the list of things to look at
[16:20]	Simon Linden:	it definitely makes sense to fix it up
[16:20]	Morgaine Dinova:	/me nods, tnx
[16:21]	Andrew Linden:	Sorta Morgain. Working on the interest list is on the backlog, but strategies about how to do it are not detailed in the backlog
[16:21]	Andrew Linden:	nevertheless, making the interestlist do 3D culling is naturally a strategy that we'll consider
[16:21]	Morgaine Dinova:	kk
[16:21]	Andrew Linden:	however, there are some other strategies that might make the 3D culling not the primary trick
[16:22]	Morgaine Dinova:	Definitely a part of FEF! to reduce downloads :P
[16:22]	Moon Metty:	hehehe
[16:22]	reddot99 Republic:	andrew, complete threorical question, can a sim border itself?
[16:22]	Andrew Linden:	FEF? You'll have to expande that TLA for me.
[16:23]	Morgaine Dinova:	Andrew: OMG, you need to listen to your esteemed leader. :-) Philip defined it as Fast East Fun!
[16:23]	Morgaine Dinova:	Easy
[16:23]	Techwolf Lupindo:	Shouldn't the intrest list be done client side and the client just tell the sim what area I need objects from? For even better results, devide each area into static blocks and the client chooses what area blocks to get.
[16:23]	Andrew Linden:	No reddot99, however it might be possible, with a little bit of work, to create non-flat topologies between distant regions
[16:23]	Ardy Lay:	reddot99 wants "mouse warp" for SL Simulators. :-)
[16:23]	Morgaine Dinova:	Fast Easy Fun!
[16:23]	reddot99 Republic:	lol, sorta like in pacman if you went into the tunnels at the side
[16:23]	Simon Linden:	Techwolf - yes, that's another strategy we're considering, letting the viewer give more data about what and where it wants updates from
[16:23]	Stryker Jenkins:	will have the accronimes on the wiki for that entry
[16:24]	Cummere Mayo:	/me keeps her comments to herself abut her thoguhts of Philips definition of FEF >.<
[16:24]	Ardy Lay:	That was a typo, mouse wrap.
[16:24]	Andrew Linden:	Yes Techwolf. That is the "server announces data channels and the viewer subscribes to those it wants" model
[16:24]	Andrew Linden:	The first pass at overhauling the interestlist will probably do that, for static geometry at frist.
[16:24]	Tegg Bode:	I was going to say what about Fast West Fun too? ㋡
[16:25]	Morgaine Dinova:	Tech: the viewer really can do a better job, yes. It has the CPU power. Not really sure why it's not done that way. It fits in perfectly with caps
[16:25]	Ardy Lay:	Maybe caps are newer than the interest list.
[16:25]	Techwolf Lupindo:	Just anothere note to thow out there, seperate static object and active/phycial objects into differen lists so one list can be static and requres very little resoruces to send it.
[16:25]	reddot99 Republic:	if it put that extra work on a second core, since most systems are multicore now, arent they?
[16:26]	Techwolf Lupindo:	One client can compute its own single intreest list vs. server that has to computer inttest list for 50 avatars.
[16:26]	Simon Linden:	we do separate the static and dyanamic objects in the interest list ... there are probably spots in it that can be sped up too
[16:27]	Andrew Linden:	yes, static and dynamic objects are already handled differently on the simulator
[16:27]	Andrew Linden:	which is why we would be able to try to work on just one type initially
[16:27]	Morgaine Dinova:	Client power scales perfectly with population, 1:1, while servers clearly don't. So yeah, in the end, you need to push as much as you can down to the client if you want regions to be FEF.
[16:27]	Andrew Linden:	so as to break the project up into more managable (deliverable) bits
[16:28]	reddot99 Republic:	remember morgaine, we also dont want properitary code to be forced back into the veiwers,
[16:28]	Andrew Linden:	Oh I see. FEF = Fast Eafy Fun. Okay.
[16:28]	Moon Metty:	hehehe
[16:28]	Morgaine Dinova:	reddot: yeah, that places a limit on it
[16:28]	Techwolf Lupindo:	One item to add to the backlog, text only clients. Don't send the steady stream of object/avatar data when the client just has to ignore it.
[16:29]	Morgaine Dinova:	Hahaha ... /me watches light bulb light up above Andrew <grin>
[16:29]	Memorial Dae:	:)
[16:29]	Simon Linden:	Right, that's another good reason for the viewer to send more information about what it wants and can handle ... the server can eventually support a wider variety of viewers
[16:30]	Moon Metty:	yeah
[16:30]	Morgaine Dinova:	Simon++
[16:30]	Cummere Mayo:	ultimate fef realtime nueral interface... >.>
[16:31]	Andrew Linden:	(No Liisa, I haven't looked at the sim bombs yet.)
[16:31]	Liisa Runo:	^^
[16:31]	reddot99 Republic:	i thought ++x was more effecient than x++
[16:31]	Cummere Mayo:	anyways... with more info options for the viewers will that mean we can have a viewer that can mute out all under 18 avis? *mutters*
[16:31]	Morgaine Dinova:	And when we start using video cards for OpenCL/CUDA/PhysX, the client processing power is going to ramp up even further.
[16:32]	Moon Metty:	does a minimized viewer recieve the same amount of data from the server ?
[16:32]	Rex Cronon:	how do u know if somebody is under 18?
[16:32]	Elisha Richez:	they act like a jackass
[16:32]	reddot99 Republic:	you dont yet,
[16:32]	Andrew Linden:	Moon, I don't think the server knows anything about the minimized state of the viewer.
[16:32]	Moon Metty:	mmhmm
[16:32]	Memorial Dae:	you hijack the webcame via http on prim and check
[16:32]	Cummere Mayo:	rex i hopw we will be able to id allot of them automatically somehow when they merge us
[16:33]	Morgaine Dinova:	Moon: the data continues to be received, but it's not processed read for rendering until you de-iconize or refocus.
[16:33]	Romaq Rosher:	/me notes many over 18 well able to act far less mature than many 12 year olds he's met.
[16:33]	Morgaine Dinova:	ready*
[16:33]	Nimh20 Vandeverre:	How is LL going to keep 16-17 year olds out of the Mature/Moderate SIMS?
[16:34]	Elisha Richez:	hopefully their parents
[16:34]	Morgaine Dinova:	Don't think that's a question for Andrew/Simon :-)
[16:34]	Tegg Bode:	Might as well dump the whole Mature/PG sim idea on mainland
[16:34]	Andrew Linden:	I don't know anything about the plans for changes to age access to SL
[16:34]	Moon Metty:	i made an observation for 1.42 and the freezes
[16:34]	Simon Linden:	I'm honestly clueless about the closing of teen SL and how that might change any policy or feature. I found out about it after it was announced
[16:34]	Nimh20 Vandeverre:	Just trowing it out there Morgaine--its ok
[16:35]	Ardy Lay:	under 18 accounts have a flag in them the simulators check against region ratings and allow them to only enter PG - General regions.
[16:35]	Moon Metty:	the first time 1.42 was rolled out, everything looked the same
[16:35]	Nimh20 Vandeverre:	TY Ardy
[16:35]	Moon Metty:	around 0.3 % of freezes after 1 day
[16:35]	Moon Metty:	then there was a restart
[16:35]	Simon Linden:	we rolled out a new version last night
[16:36]	Moon Metty:	and now the average is lower, like 0.1 %
[16:36]	Elisha Richez:	i noticed
[16:36]	Andrew Linden:	there was a bug in 1.42 that caused twisted prims to have the wrong collision shape
[16:36]	Andrew Linden:	(my fault)
[16:36]	Panel 3:	Freeze of 0.814248s detected (Mono)
[16:36]	Panel 3:	Freeze of 1.023485s detected (LSO)
[16:36]	Moon Metty:	so there was no change in any code that could affect the freezes?
[16:36]	Andrew Linden:	What are these "freezes" you're talking about Moon?
[16:36]	Moon Metty:	well, the tp/Mono freezes
[16:36]	Tegg Bode:	It takes a truely great man to admit their fault (or a whole lot of people behind him with knives ㋡ )
[16:37]	Moon Metty:	the sim stalls
[16:37]	Andrew Linden:	You're saying the frezzed are less frequent in 1.42?
[16:37]	Moon Metty:	svc-4196 and all that
[16:37]	Avatar entering	sim or rezzing object causes sim to freeze for up to 30 seconds - everything stops for everybody there
[16:37]	Moon Metty:	it's too early to tell Andrew
[16:37]	Rex Cronon:	how spheres with dimples? they have different collision shape too?
[16:37]	Moon Metty:	but after the second restart, they look lower
[16:37]	Andrew Linden:	I don't think the fixes for mono got into 1.42
[16:38]	Andrew Linden:	do you know the status of those Simon?
[16:38]	Simon Linden:	they aren't in yet
[16:38]	Moon Metty:	then it's pure coincidence :)
[16:38]	Andrew Linden:	I think the code is done but just needs some test-deploy and QA attention
[16:38]	Simon Linden:	Kelly's working on it, and I'm not sure when it's due
[16:38]	Andrew Linden:	ok
[16:38]	Moon Metty:	the freezes are always short after a restart
[16:38]	Moon Metty:	apparently there's some fluctuation
[16:39]	Andrew Linden:	There was some good stuff in 1.42. I can't remember any of it 8-0
[16:39]	Simon Linden:	Our code release pipeline is in flux too ... we're trying to move to methods that will have more frequent updates, so it cuts down the time between fixing and releasing
[16:39]	Nimh20 Vandeverre:	I thought the MONO bug was this JIRA
[16:39]	Nimh20 Vandeverre:	http://jira.secondlife.com/browse/SVC-3895
[16:39]	Panel 3:	Freeze of 1.287215s detected (LSO)
[16:39]	Panel 3:	Freeze of 1.290181s detected (Mono)
[16:39]	Rezzing Mono	scripted object cripples sim FPS
[16:39]	Stryker Jenkins:	when is the rest of the roleout for 1.42 planed by the way? I seem to miss a blog posting about that ...
[16:39]	Simon Linden:	that meeter seems to cause freezes :)
[16:39]	Stryker Jenkins:	only saw the Pilot announcement
[16:39]	Simon Linden:	should be next week Wednesday
[16:40]	Andrew Linden:	looks like -4196 is a duplicate of -3895
[16:40]	reddot99 Republic:	i couldnt find release nots on it either
[16:40]	Panel 3:	Freeze of 0.585719s detected (Mono)
[16:40]	Moon Metty:	lol Simon
[16:40]	Panel 3:	Freeze of 0.569920s detected (LSO)
[16:40]	Panel 3:	Freeze of 0.734183s detected (Mono)
[16:40]	Panel 3:	Freeze of 0.814019s detected (Mono)
[16:40]	Panel 3:	Freeze of 0.602572s detected (LSO)
[16:40]	Ardy Lay:	Close either of those as "duplicate" at your own peril.
[16:41]	Ardy Lay:	I can see the mobs forming already.
[16:41]	reddot99 Republic:	lol
[16:41]	Moon Metty:	there are a few in SEC too
[16:41]	Cummere Mayo:	/me hides the pitchfork
[16:41]	reddot99 Republic:	yeah, i think an old sec i submited can be attributed to that
[16:41]	Moon Metty:	best is to leave them be indeed, Ardy
[16:42]	Moon Metty:	hehe
[16:42]	Stryker Jenkins:	duty calles... see you all again some day
[16:42]	Stryker Jenkins:	Bye Bye!
[16:42]	Rex Cronon:	tc
[16:42]	MoContur Chastity	Belt: whispers: MoPlugin Cloth Undress: Undress: Undershirt
[16:42]	Moon Metty:	bye Stryker
[16:42]	reddot99 Republic:	oh hey, my jira was 4 months older than either of those
[16:43]	reddot99 Republic:	lol
[16:43]	Romaq Rosher:	I've a silly question... would anyone have a sculpty that shows exactly where the sculpty collsion shape surface is?
[16:43]	Romaq Rosher:	Or is this something that Andrew could generate for us out of the collision data he has access to from the server?
[16:43]	reddot99 Republic:	http://jira.secondlife.com/browse/SEC-154
[16:43]	Andrew Linden:	here is how you figure out what the collision shape of a sculpty is...
[16:43]	Liisa Runo:	just toggle the sculpt to sphere and you see the collision shape
[16:43]	Andrew Linden:	turn it into a torus, then pretend you took the convex hull of that
[16:43]	reddot99 Republic:	really old sec that shows the same bug as the jira
[16:44]	reddot99 Republic:	erm. as the mono object one
[16:45]	MoContur Chastity	Belt: whispers: MoPlugin Cloth Undress: Undress: Undershirt
[16:45]	Romaq Rosher:	If that's true, then the orientation in 3d of your mesh could... well, mean a great deal.
[16:45]	Romaq Rosher:	As far as collision, anyway.
[16:46]	Simon Linden:	Romaq - yeah, the physics representation of sculpties isn't very accurate
[16:46]	Romaq Rosher:	Thank you, Andrew. That's very simple, effective, and something I can test. I don't mind that it's not accurate, accuracy would be terribly expensive.
[16:46]	Simon Linden:	That should get better with the mesh work, or at least there will be alternatives
[16:46]	reddot99 Republic:	yeah, but its better to make the sculpt collisionless, and give a second second prim for actuall collision detection
[16:46]	Romaq Rosher:	Having a means to judge what it *is* is terribly useful.
[16:47]	Simon Linden:	right, actually being able to view collision shapes would really help
[16:47]	Romaq Rosher:	The server/ client interaction knows 'something' about the collision and where it should happen. I'm just looking for a means to expose what that interaction is visually so I can plan meshes around it.
[16:47]	Simon Linden:	yep, ANdrew just said he was crashing
[16:47]	reddot99 Republic:	could there be a mod for a veiwer to 'see' nothing but collision skeletons?
[16:48]	Simon Linden:	Currently the collision info is never sent to the viewer, so no
[16:48]	Romaq Rosher:	Yeah, we tinys have an odd collision mess because even though I'm folded, the client still treats me as a very small 2-legged standing avatar.
[16:48]	reddot99 Republic:	i mean, could the collision info be sent
[16:48]	Romaq Rosher:	mess=mesh
[16:48]	Simon Linden:	Falcon has done some work in this area, and I know that's being considered
[16:48]	reddot99 Republic:	it a mess as well, romaq
[16:48]	Romaq Rosher:	Aye. :(
[16:49]	Simon Linden:	We've had some working prototypes on using prims to specify the shape, or a simplication such as "use a box for this"
[16:49]	reddot99 Republic:	though, i finally got something that might work to imitate an avatar, but makes a collision skeleton smaller than normal
[16:50]	Rex Cronon:	what happened with those prototypes, simon:)
[16:50]	Simon Linden:	We don't want to penalize folks for doing the right thing, so (for example) special prims that are just used for shapes shouldn't count againt prim limits
[16:51]	Romaq Rosher:	:)
[16:51]	Simon Linden:	This is all being gathered together in the mesh project since that has the exact same problem - complex visual shapes that would be too expensive for the physics engine
[16:51]	Moon Metty:	welcome back Andrew :)
[16:51]	Andrew Linden:	apparently my linux machine was getting pretty hot
[16:51]	Romaq Rosher:	I recall Philippe saying how mesh, for example, could replace the 200 prim hair on avatars. Is that to say mesh is planed with flexiblity, as with flexi-prims?
[16:52]	Andrew Linden:	but it didn't hard-reboot this time
[16:52]	Romaq Rosher:	Philip, sorry..
[16:52]	Cummere Mayo:	ooh :-/ time to get a better cooling sys?
[16:52]	s corset	lace (L): whispers: Psi loosens the laces on her corset
[16:52]	reddot99 Republic:	but wouldnt it be simplest to give the physics engines more power to work with?
[16:52]	Andrew Linden:	well, this is a MacBook pro, and I've got it on some fans
[16:53]	Romaq Rosher:	Power is additional expense. More physics calculations have to take up CPU somewhere, and if it's on the client, the client is known for lying. Early and often.
[16:53]	Cummere Mayo:	thats your problem right there... *hates macs*
[16:54]	reddot99 Republic:	to be consistent, it'd have to be server side
[16:54]	Romaq Rosher:	Aye. I shot you dead. I told my client to say so.
[16:54]	Andrew Linden:	I don't think "animated" mesh would be part of the initial feature set
[16:54]	Andrew Linden:	so I don't think mesh could replace flexi hair
[16:54]	Andrew Linden:	initially
[16:54]	Romaq Rosher:	Ugh.
[16:54]	Morgaine Dinova:	TPVs will animate it.
[16:55]	Andrew Linden:	mac hardware is pretty good. this laptop has been a trooper for three years
[16:55]	Romaq Rosher:	Not having flexi as part of mesh rollout will greatly slow the adoption for replacements to prim hair.
[16:55]	reddot99 Republic:	i've got my old ibook still running clean like 6 years on
[16:55]	Moon Metty:	there will be enough to replace, Romaq
[16:55]	Moon Metty:	it's good it's not all in one go
[16:55]	Romaq Rosher:	Well, all the prims on *ME* would be a good start.
[16:56]	reddot99 Republic:	i'd be a bigger challenge, i think, romaq
[16:56]	Romaq Rosher:	I've spent evenings trying to decide to make my own sculpty ferret vs. stabbing my eyes out with forks. Decisions, decisions.... :(
[16:56]	Moon Metty:	hehehe
[16:56]	Romaq Rosher:	True, RedDot. :)
[16:57]	Romaq Rosher:	Yeah, you're pretty much all prim there.
[16:57]	reddot99 Republic:	mostly sculpts
[16:57]	Romaq Rosher:	Yeah... all the sculpty prims just to make custom body avatars. That can't do anything useful for bandwidth.
[16:58]	reddot99 Republic:	also, question, has havok 7 reduced avatar walk and run speeds?
[16:58]	Romaq Rosher:	But I don't see myself on SL as 'human form'. I just don't. What you see is how I think of me here.
[16:58]	Simon Linden:	reddot - I don' t think the limits changed
[16:59]	reddot99 Republic:	it used to be 3.5 walk, and 7 m/s run
[16:59]	reddot99 Republic:	now runs like 5 something m/s
[16:59]	reddot99 Republic:	and walk is 3.1
[16:59]	Andrew Linden:	really? hrm.
[16:59]	Andrew Linden:	It shouldn't have changed.
[17:00]	Manji Nishizawa:	I feel that friction grew big in all material
[17:00]	Andrew Linden:	However it could have. I had plenty of trouble trying to match the speeds in 1.8 --> 4.6
[17:00]	Cummere Mayo:	ive noticed it as well but i like it. slightly more realistic
[17:00]	Moon Metty:	and flying?
[17:01]	reddot99 Republic:	i dont know what default flight speed used to be, only walk and run
[17:01]	Moon Metty:	i always have trouble flying because it's too fast lol
[17:01]	reddot99 Republic:	though, velocity change for walk and run is consistent at least
[17:02]	Andrew Linden:	Simon could you grab the OH transcript for me?
[17:02]	Simon Linden:	sure
[17:02]	Andrew Linden:	Ok, I've got to run off.
[17:02]	Andrew Linden:	Have a good weekend everyone.
[17:02]	Rex Cronon:	tc andrew
[17:02]	reddot99 Republic:	lsl says walk speed is now 3.2
[17:02]	Andrew Linden:	Thanks for coming.
[17:02]	Moon Metty:	thanks for the meeting :)
[17:02]	Romaq Rosher:	Thank you for your time, ANdrew!
[17:03]	Psi Merlin:	Thanks Andrew
[17:03]	Romaq Rosher:	And Simon!
[17:03]	reddot99 Republic:	run is 5.13
[17:03]	Moon Metty:	see you soon Andrew
[17:03]	Tegg Bode:	Yea flying always sem to fast, there should probably be 2 fly speeds like there are walk/run
[17:03]	Simon Linden:	thanks everyone ... see you next time
[17:03]	Simon Linden:	bye
[17:03]	Manji Nishizawa:	see you
[17:03]	Moon Metty:	bye Simon :)
[17:03]	Tegg Bode:	Cyas ㋡
[17:03]	Moon Metty:	good idea Tegg
[17:03]	Rex Cronon:	tc simon

Generated with SLog Wikifier