User:Andrew Linden/Office Hours/2008 05 17

Transcript of Andrew Linden's office hours:

[11:01]	Creem Pye:	howdy
[11:01]	Jrift Sol:	hi
[11:01]	Andrew Linden:	hello
[11:01]	Yuu Nakamichi:	ao off
[11:03]	Andrew Linden:	I've got a Resident asking for "sleep and hungry" as features in SL
[11:03]	Andrew Linden:	What do you all think of that? ;-)
[11:03]	Jrift Sol:	omg!
[11:03]	Rex Cronon:	no
[11:03]	Rex Cronon:	i don't want to eat prims:)
[11:03]	Talarus Luan:	As long as I can eat other avatars to feed a Dragon's hunger, OK.
[11:03]	Seifert Surface:	that one has come up regularly since time began
[11:03]	Rex Cronon:	is sl becoming an rpg?
[11:04]	Talarus Luan:	Most ESPECIALLY griefers and land extortionists. :)
[11:04]	Talarus Luan:	They should be a meal unto themselves.
[11:04]	Dafydd Laville:	I'd not mind it as an option. I think it would be popular amongst the roleplayers.
[11:04]	Kitto Flora:	What I think is fix the bugs before adding new features
[11:04]	Dafydd Laville:	Agreed, Kitto
[11:04]	Seifert Surface:	people can script something like food if they want
[11:04]	Creem Pye:	heh you could add artificial latency to simulate drowsiness
[11:04]	Yuu Nakamichi:	yeah.. this woudl be pretty low on my list as well
[11:05]	Andrew Linden:	What if it were opt-in... who here would opt in for requiring food, sleep, and other mundane stuff?
[11:05]	Seifert Surface:	how about a third of time you need to be asleep too
[11:05]	Creem Pye:	and maybe a fullscreen black HUD that gradually gets more opaque as the avatar passes out
[11:05]	Yuu Nakamichi:	Can I haz a portal gun, Andrew?
[11:05]	Seifert Surface:	your screen is black, oh wait, there's an excuse for regular downtime
[11:05]	Jrift Sol:	that kind of stuff belongs in resident creations, not SL itself, IMHO
[11:05]	Talarus Luan:	Well, next, you will be looking to add opt-in for pregnancy. O.o
[11:05]	Kitto Flora:	I'm too busy scripting around and explaining built-in long-time bugs to do any eat or sleep.
[11:05]	Arawn Spitteler:	We already of a Lag Equivalent to Meteorology.
[11:06]	Andrew Linden:	interestingly... there are some render/visual techniques that can be done to degrade the visual quality of SL that may actually provide more immersiveness
[11:06]	Yuu Nakamichi:	yeah how about a grid weather service
[11:06]	Yuu Nakamichi:	such as?
[11:06]	Dafydd Laville:	I would opt in. Then script an auto-feed ;)
[11:06]	Arawn Spitteler:	Pregnancy happens, and food that dissappears as eaten would also be nice.
[11:06]	Andrew Linden:	the eye is not a perfect visual tool and has slow response to sudden light changes, for example
[11:06]	Creem Pye:	ahh HDR type stuff?
[11:06]	Andrew Linden:	and does poorly at distinguishing colors at low light levels
[11:07]	Jrift Sol:	a lot of visual tricks can also be done via a hud that can cover the entire screen
[11:07]	Andrew Linden:	Some games use this for immersive enhancement.
[11:07]	Creem Pye:	those effects woudl be great to have, but I guess that would depend on the viewer team almost entirely?
[11:08]	Andrew Linden:	Right, but how many here would opt-in to such visual effects if they were checkbox options in the preferences?
[11:08]	Creem Pye:	/me would
[11:08]	Seifert Surface:	maybe, would have to try it out
[11:08]	Creem Pye:	shadows too of course =)
[11:08]	Kitto Flora:	I doubt it.. no.
[11:08]	Talarus Luan:	Given that I can barely run with just the basic shaders in 1.19.1.4, I'd say it would most likely remain OFF for me.
[11:08]	Jrift Sol:	wouldn't. i code for 24 hours straight sometimes. i don't want my avatar running out of juice before my rl body
[11:08]	Yuu Nakamichi:	it could easily look plasticky. if you know what I mean
[11:08]	Kitto Flora:	Ah - shadows - that would be good. Real shadows.
[11:08]	Rex Cronon:	hi again. my network just crashed:(
[11:08]	Andrew Linden:	Ah well, I was just curious. Lessee... what is the news?
[11:08]	Yuu Nakamichi:	the problem with ehanced immersiveness is the
[11:09]	Yuu Nakamichi:	uncanny valley
[11:09]	Talarus Luan:	News?
[11:09]	Talarus Luan:	You tell us. :P
[11:09]	Dafydd Laville:	I think most would try it to see what it is like. Then turn it off if it was slightly inconvenient.
[11:09]	Arawn Spitteler:	/me wouldn't, but can see how it might be required for an RPG: Didn't Boston Legal do a story on a kid who played video games to death? Was there a Real world incident, being relayed?
[11:09]	Yuu Nakamichi:	or if it wasn't entirley believable so to speak
[11:09]	Andrew Linden:	Yeah, I was just thinking out loud... I don't have many announcements to make.
[11:09]	Andrew Linden:	I recently closed a bug about the "Swimmer 1.1" for a "won't finish".
[11:10]	Jrift Sol:	i think what would be more useful for rpg is to allow estate owners to disable parts of the client, such as name tags and the minimap
[11:10]	Jrift Sol:	the rest could be done in lsl
[11:10]	Andrew Linden:	Turns out the attachment's script relies on a quirk in the Havok1 system that causes contant moving_start() and moving_end() events for an avatar in swim mode.... even when the avatar is moving smoothly.
[11:10]	Kitto Flora:	Is the Swimmer 1.1 maker providing a fixed upgrade?
[11:10]	Arawn Spitteler:	Maybe Clients that report that such features are disabled.
[11:11]	Andrew Linden:	No Kitto. Supposedly Siggy Romulus made it and is no longer active in SL
[11:11]	Andrew Linden:	However some later versions of the swimmer work ok...
[11:11]	Kitto Flora:	OK
[11:11]	Rex Cronon:	jrift, i don't think u want that. what i think it would be better is if there were scripts that chould check if those things are on or not
[11:11]	Rex Cronon:	i mean lsl functions
[11:11]	Jrift Sol:	/me agrees, rex. good idea
[11:11]	Kitto Flora:	Theres a communal swimmer object out now. Dont need an attachment
[11:12]	Andrew Linden:	Hrm... there was something I wanted to bring up in these office hours but I can't remember what it was.
[11:12]	Andrew Linden:	Perhaps I'll remember before the end.
[11:13]	Dafydd Laville:	IS the communal swimmer the one that rezzes buoys to sit on?
[11:13]	Kitto Flora:	Av vertical drift? (when flying)
[11:13]	Andrew Linden:	So I'm going to just open up the table for topics. I don't have much on the agenda today.
[11:13]	Talarus Luan:	Well, I have a topic for more discussion... megaprims. :)
[11:13]	Arawn Spitteler:	Did you write it down, ten minutes after the last meeting, someplace you'd surely never forget?
[11:13]	Andrew Linden:	Ah vertical drift.... I fixed it for the "drift down" problem.
[11:13]	Yuu Nakamichi:	oh?
[11:13]	Trent Platthy:	Yay
[11:13]	Andrew Linden:	That fix may be in the next update to the server (1.22)
[11:14]	Kitto Flora:	The 'communal swimmer' is an object that you click, and it makes you swim in water. Rather like a dance ball I think.
[11:14]	Talarus Luan:	That's good. was getting old to be hovering and five minutes later hovering on the ground
[11:14]	Arawn Spitteler:	/me was just sliding at CG's, so clicked the Fly Button, and took to bobbing up and down.
[11:14]	Andrew Linden:	There was some other (non-havok4) related problem that may have held up the next server update this week
[11:14]	Andrew Linden:	someone was going to make a decision about it this morning
[11:14]	Andrew Linden:	and I haven't heard the latest
[11:14]	Jrift Sol:	i'd like to raise issue with push being so severly dampened over what is a very short distance (though i imagine that's been discussed a lot already)
[11:14]	Andrew Linden:	If they decide to go forward I think the update rollout would happen this week... starting Wednesday.
[11:15]	Talarus Luan:	As far as I can tell, llPushObject doesn't even work anymore
[11:15]	Jrift Sol:	it seems nearly impossible to push an avatar horizontally
[11:15]	Jrift Sol:	it works, but only if the pusher is basically in contact with the target
[11:15]	Andrew Linden:	All of the fixes I've been talking about for the last two or three weeks are waiting for the 1.22 update.
[11:15]	Trent Platthy:	It does, they've just turned the falloff WAAAY up.
[11:15]	Yuu Nakamichi:	what was the cause for the drift down?
[11:15]	Andrew Linden:	I thought Kelly Linden fixxored the llPushObject() ineffectiveness against avatars.
[11:16]	Andrew Linden:	I wonder where that fix was checked in...
[11:16]	Trent Platthy:	Nope, still seems to be fuxxored.
[11:16]	Jrift Sol:	maybe that's in 1.22
[11:16]	Seifert Surface:	jrift: is this pushing an av that is in contact with the ground?
[11:16]	Andrew Linden:	I'll ask kelly in another channel...
[11:16]	Jrift Sol:	yes
[11:16]	Jrift Sol:	to get any kind of movement, i have to push up, then horizontally. this works, but often the av enters falling state, which locks the position temporarily
[11:16]	Seifert Surface:	my pet problem is with pushing avatars that are in freefall, but yeah there has always been some friction with the ground
[11:17]	Rex Cronon:	could at least objects that u own be allowed to push u harder?
[11:17]	Kitto Flora:	What you using llPushObject() on Avs for Jrift<
[11:18]	Jrift Sol:	i make weapons systems for use in push enabled areas. it might sound 'dodgy', but it's a valid market
[11:18]	Andrew Linden:	It seems that Kelly did checkin some fixes for the llPushObject() problem against avatars and it should be in 1.22 Server.
[11:18]	Kitto Flora:	Sure. OK
[11:18]	Yuu Nakamichi:	any news on the avatar bounding box issue - https://jira.secondlife.com/browse/SVC-2014
[11:18]	Yuu Nakamichi:	"pending"?
[11:18]	Yuu Nakamichi:	ah, I see your comment
[11:19]	Yuu Nakamichi:	great
[11:19]	Andrew Linden:	Yuu, I fixed that problem but it just missed the 1.22 deadline.
[11:19]	Jrift Sol:	i look forward to exploring 1.22
[11:19]	Yuu Nakamichi:	ok
[11:19]	Andrew Linden:	Uh... that is SVC-2014.
[11:19]	Andrew Linden:	BTW, if I didn't ask your question try asking again. I try to scroll back but don't always catch everything.
[11:20]	Kitto Flora:	I've received some complaints about the force of ejection of tall Avatars that become embedded in relativly small prims.
[11:20]	Andrew Linden:	on standing up from a seat Kitto?
[11:20]	Andrew Linden:	That is the Havok4 penetration resolver at work.
[11:20]	Kitto Flora:	Yes, stand up and intersect a prim, which then tosses them way up in the air
[11:21]	Andrew Linden:	I tried to fix the "avatar dismount" code but ended up breaking a bunch of content so it was rolled back.
[11:21]	Andrew Linden:	I want to take another swing at that one.
[11:21]	Kitto Flora:	Lots higher for tall prims. My size - its not a big effect
[11:21]	Andrew Linden:	However... that isn't yet on my short list of things I need to work on.
[11:22]	Kitto Flora:	Well its not particularly a dismont problem
[11:22]	Andrew Linden:	I mostly get my short list from Sidewinder these days, but bugs mentioned here may also get my attention.
[11:22]	Creem Pye:	how about an oldie but goodie? SVC-22: Parcel restrictions inappropriately applied to vehicles on border crossings.
[11:22]	Andrew Linden:	Not dismount Kitto? You mean on teleport into penetration?
[11:22]	Talarus Luan:	Well, I don't have a question so much as making an appeal to reason to reinstate megaprim creation.
[11:23]	Andrew Linden:	/me has a plan for megaprim liberation.
[11:23]	Talarus Luan:	I am aware. :)
[11:23]	Trent Platthy:	Hehe, liberation. Free the prims! :)
[11:23]	Andrew Linden:	Simply reopening all prim sizes is not part of the plan.
[11:23]	Jrift Sol:	would it be possible to shut down physics for objects based on the number of intersections caused by a particular avatar? this would go a long way towards preventing the types of attacks that the old griefer known as 'dench' uses?
[11:23]	Kitto Flora:	It apears that the problem is the force of ejection of a prim-embedded avatar. It so happens that the embed in the presented case is due to a dismount.
[11:23]	Rex Cronon:	that i kind of a long term plan
[11:24]	Talarus Luan:	Well, the horse is long gone. Closing the barn door at this point has about the same net effect as intrusive copyright protection mechanisms.
[11:24]	Andrew Linden:	Jrift, describe dench's grief method.
[11:24]	Jrift Sol:	he leaves multiple large object rezzers around a sim
[11:25]	Trent Platthy:	Thats actually a good idea. Keep a running total of the number of collisions caused by objects owned by a key over the last X minutes, and if they exceed a threshold, kill physics for his objects.
[11:25]	Jrift Sol:	these spew out 10m3 cubes at such a rate that intersections go through the roof
[11:25]	Talarus Luan:	You are annoying and limiting legitimate, legal, proper use for the supposed benefit of stopping illegal improper use.
[11:25]	Jrift Sol:	at the same time, he often leaves many twisted physical torii around the sim
[11:25]	Trent Platthy:	He also like leaving around rezzers that rez hundreds of freebie houses, deeply nested. Which tends to kill the sim. :)
[11:25]	Andrew Linden:	There is already code in place that will disable collisions between an avatar and another object if the avatar is penetrating deep enough (1 meter) , for long enough (1 second).
[11:26]	Talarus Luan:	yes. That isn't a valid argument against megaprims in itself.
[11:26]	Andrew Linden:	I made some adjustments to that in 1.22... I reduced the distance, and added some other logic
[11:26]	Talarus Luan:	For many reasons, even.
[11:26]	Jrift Sol:	these are object to object collisions causing the problems
[11:26]	Andrew Linden:	which should help for the cases where the avatar is stuck in smaller objects, or in an array of smaller objectgs.
[11:26]	Rex Cronon:	and very fast rezzing
[11:27]	Talarus Luan:	1) Megaprims are already in existence, and rezzers can happily rez away 1024x1024x1024 cubes to their heart's content.
[11:27]	Andrew Linden:	Trent, these nested houses kill the Havok4 sim?
[11:27]	Creem Pye:	is there a way to limit H4's penetration avoiding force on avatars?
[11:27]	Creem Pye:	(which I guess causes the avatars to be pushed hard when intersecting with prims)
[11:27]	Trent Platthy:	They don't crash it that I've seen, but they do leave it at a TD of 0.01.
[11:27]	Andrew Linden:	Creem, could you rephrase that?
[11:28]	Jrift Sol:	the sims often don't fall all the way over, but are often dragged down to 0.01, which is effectively the same thing to try and work in
[11:28]	Andrew Linden:	Should there be a limit to how big a dynamic object can be?
[11:28]	Andrew Linden:	What if I forced all obejcts larger than some size to be static?
[11:28]	Creem Pye:	well this force we see when dismounting, it seems that the force is proportional to how deeply embedded in the object the avatar is. perhaps if the force was capped toa certain threshold, the effect would be less obnoxious and griefer-friendly
[11:28]	Trent Platthy:	Its not so much size as rezzing hundreds of them.
[11:28]	Scalar Tardis:	I wanted to ask if there's any punitive action against the creation of non-griefing megaprims before there's formal plan in place for them.
[11:28]	Trent Platthy:	Then he'd just use smaller prims.
[11:28]	Talarus Luan:	Yes.
[11:29]	Jrift Sol:	i'm not sure that'd do it. the sim could be lagged just as badly by rezzing hundreds of 0.5m3 cubes in an enclosed space at high speed
[11:29]	Talarus Luan:	Again, not an argument against megaprims.
[11:29]	Trent Platthy:	5000 1m^3 prims will kill the sim just the same as 5000 10m^3 prims.
[11:29]	Talarus Luan:	Griefers will grief.
[11:29]	Scalar Tardis:	well 5000 tiny prims only if touching and all rezzed at the same time :)
[11:29]	Jrift Sol:	i'm not arguing against megaprims. just trying to figure a way of limiting griefer's impact on sims by abusing physics
[11:29]	Andrew Linden:	No, I meant... perhaps objects that are larger than... say 256 meters should be forced static. Anyone here have a problem with that?
[11:30]	Andrew Linden:	Perhaps it should be 128?
[11:30]	Talarus Luan:	Why is it an issue, Andrew?
[11:30]	Rex Cronon:	yes
[11:30]	Creem Pye:	larger than 256m and physical... tht's hard to imagine for legit use
[11:30]	Seifert Surface:	256 i cant imagine there would be an objection
[11:30]	Talarus Luan:	What problem are you trying to solve?
[11:30]	Arawn Spitteler:	Someone might want to build a yacht
[11:30]	Creem Pye:	Gaius might object if you limited physical objects to 32m though ;)
[11:30]	Seifert Surface:	128 there might be uses
[11:30]	Rex Cronon:	size not the problem
[11:30]	Jrift Sol:	i'm trying to solve the problem of large amounts of physical intersections caused by one person being able to affect a region
[11:31]	Andrew Linden:	Grief modes Telarus. Someone was just complaining about piles of giant dynamic megaprims I think.
[11:31]	Scalar Tardis:	well I intend to explore these megaprim uses for physics. Non-griefing, of course. :)
[11:31]	Trent Platthy:	They aren't megaprims. Not usually anyway.
[11:31]	Jrift Sol:	it seems the megaprim and physics discussions kind of got merged :)
[11:31]	Jrift Sol:	no, they're normally 10m3
[11:31]	Talarus Luan:	Yes.
[11:31]	Andrew Linden:	well Jrift, the problem there is making the simulator smart enough to spot the bad from the good.
[11:31]	Trent Platthy:	Only thing I've seen him use megaprims for is using the 'infinity plane' one with an annoying moire pattern.
[11:31]	Scalar Tardis:	well I don't like the word "mega" anyway. because under 256 m it's not really mega at all. 1000m is mega. 25m is not.
[11:32]	Andrew Linden:	There is already code that tries to keep the simulator alive in extreme physics load.
[11:32]	Jrift Sol:	that's why i was thinking about intersections per agent. i have no idea if that's practical or even possible though
[11:32]	Andrew Linden:	That is the primary reason why Havok4 does not crash so easily.
[11:32]	Andrew Linden:	I've talked about the RCCS (runtime collision control system) in the past
[11:32]	Jrift Sol:	does that system have any concept of object owner?
[11:33]	Andrew Linden:	and it is the primary thing standing between you and crippling lag in the simulator under lots of dynamic objects.
[11:33]	Trent Platthy:	It doesn't do that standing very well though.
[11:33]	Andrew Linden:	No Jrift, it does not know about object owner yet.
[11:33]	Rex Cronon:	the problems is that griefers rezz thousands of physical prims in very short time, which lags a sim to death
[11:33]	Trent Platthy:	Well, its not so much the rezzing as them collidiing.
[11:34]	Talarus Luan:	Again, I understand the physics issue, but it is an overall issue, not directly related or solely caused by megaprims.
[11:34]	Andrew Linden:	Ok, so you all would be in favor of clamping down on dyanmic objects on a per-owner basis.
[11:34]	Kitto Flora:	Have griefers found some way past the grey-goo fence, that enableles them to rez these thousands of objects?
[11:34]	Scalar Tardis:	well that and object return would be nice to have a priority of what gets shut down, going from visitor, to group, to land owner, to estate owner.
[11:34]	Creem Pye:	so you're asking for another type of grey goo mitigator?
[11:34]	Jrift Sol:	yes
[11:34]	Rex Cronon:	if they are physical is kind of obvious they collide too:)
[11:34]	Trent Platthy:	Talarus, none of us are talking about megaprims.
[11:34]	Andrew Linden:	What kind of clamping down are you thinking? Pinning such objects in place? Deleting them?
[11:34]	Trent Platthy:	Pin them. Kill physics interaction for them
[11:34]	Jrift Sol:	turning off physics would save the sim
[11:34]	Jrift Sol:	object return would clean up later
[11:35]	Rex Cronon:	how about return them to owner?
[11:35]	Andrew Linden:	Kitto, you can rez thousands under the grey goo fence if you rezz them slow enough (but still fast)
[11:35]	Talarus Luan:	Basically, what this boils down to is this: I am a responsible user of megaprims. I use them creatively and to reduce primitive load in sims (positive, yes?). I use them aesthetically (positive, yes?).
[11:35]	Andrew Linden:	the rate would be... about 320 per second or lower I think
[11:35]	Jrift Sol:	i'm concerned about returning to owner. that's 1000s of asset transactions
[11:36]	Talarus Luan:	As such, I want to get exacting fits, but I can't do that with depending on even the THOUSANDS of megaprims that are in existence out there right now.
[11:36]	Trent Platthy:	o worse than them getting shoved off the sim border.
[11:36]	Trent Platthy:	*No
[11:36]	Arawn Spitteler:	Wouldn't the original rezzing be an asset transaction?
[11:36]	Andrew Linden:	Well Jrift, we already batch up object returns, so not necessarily 1000's of assets returned, but 100's perhaps.
[11:36]	Jrift Sol:	ok. i'll shut up for a bit to allow talarus to voice his thoughts on megaprims
[11:37]	Scalar Tardis:	it would ultimately be best if a visitor spamming a sim only ends up spamming themselves. object creation rights should have some say in "physics CPU time usage"
[11:37]	Rex Cronon:	if is a copy of one object, do u even need to return it?
[11:37]	Talarus Luan:	I was VERY happy to see a way to create custom >10m3 prims, and was planning on finishing up my mall the way I envisioned, instead of tiling hundreds of 10m3 prims.
[11:37]	Rex Cronon:	i think that megaprims do help reduce lag
[11:38]	Talarus Luan:	I know that it has been planned, and a long time in coming. I even remember Sidewinder telling us that the limits were off, so I know that it was intentional, not a bug.
[11:38]	Seifert Surface:	talarus: what are you asking for?
[11:38]	Talarus Luan:	I've known since H4 was in late beta that the prim creation limits were off, and was waiting for LL to just turn it off in the client.
[11:39]	Talarus Luan:	Simple. Turn them back on, please. Limit them to 256m3, and let us get to adding net positive content back to the grid.
[11:39]	Andrew Linden:	No Telarus, the bug that allowed new megaprims was not intentional. What Sidewinder was saying was that "Andrew Linden's plan to clamp all megaprims >256 to 256 has been cancelled."
[11:39]	Creem Pye:	btw has anybody done methodical on the effect of large linked objects (made of megaprims or small prims) on physics time?
[11:39]	Rex Cronon:	u need a empty sim for that creem
[11:39]	Creem Pye:	like for floors, etc... I'm wondering how the collision islands are panning out in H4
[11:39]	Talarus Luan:	Andrew, I am talking about H4 beta meetings, and he DID say what I just said. I can spend some time and dig up the specific quotes if you like, but I do remember it being said.
[11:39]	Andrew Linden:	We're going to have a meeting in LL about megaprims on Thursday, right before my Thursday hours.
[11:40]	Andrew Linden:	Well, I don't believe it Telarus, so you'll have to dig it up for me.
[11:40]	Talarus Luan:	Otherwise, why would I have the expectation and knowledge I did, for a couple months now?
[11:41]	Talarus Luan:	Very well. :)
[11:41]	Andrew Linden:	And if he did say that megaprims had been opened wide on purpose then he was mistaken.
[11:41]	Gaius Goodliffe:	I will note (as I frequently do) that there are no parcel encroachment problems or sim border issues with smaller megaprims that we don't already have without them (due to linksets). Up to the size of a linkset, they don't create any new problems. We really ought to just allow individual prims to be up to the size of a linkset.
[11:41]	Dafydd Laville:	Excuse my ignorance but is the issue only with dynamic megaprims?
[11:42]	Talarus Luan:	Gaius: absolutely
[11:42]	Talarus Luan:	Though I don't see a reason to limit them less than the size that wouldn't encroach OUTside of a sim.
[11:42]	Talarus Luan:	IE >256m in any dimension
[11:42]	Andrew Linden:	Not true Gaius. The way megaprims are considered for linkability for legacy objects means that with megaprims you can create new objects that exceed the normal 54 metere bounding sphere.
[11:42]	Jrift Sol:	gaius, i think that's the root of the problem - that objects can overhang parcels despite restrictions
[11:43]	Rex Cronon:	not only megaprims
[11:43]	Gaius Goodliffe:	That really don't help much for making objects much larger than a linkset. It only allows you to link over linkset sized prims into one.
[11:43]	Andrew Linden:	No Dafydd, the problem is not with dynamic megaprims only.
[11:43]	Seifert Surface:	maybe the issue is that nobody wants to spend time coding a stopgap measure which will go obselete when the real solution is there, when they could be coding the real solution?
[11:43]	Talarus Luan:	Objects megaprims or not already encroach. Again, not a valid argument against megaprims.
[11:43]	Dafydd Laville:	Prim bridging sim borders?
[11:44]	Andrew Linden:	The main problem with megaprims is the "prim encroachment" problem... objects sticking over onto your parcel from your neighbors.
[11:44]	Gaius Goodliffe:	(Try making a large object with megaprims smaller than a linkset, and you won't get anything much larger than you can get without them.)
[11:44]	Talarus Luan:	That is not a problem endemic to megaprims, Andrew
[11:44]	Rex Cronon:	is already possible to make a linked set using regular prims that overhang over someobyd elses property
[11:44]	Morgaine Dinova:	As Talurus said, that's not specific to megaprims
[11:44]	Andrew Linden:	Correct Telarus, but megaprims exacerbate the problem.
[11:44]	Talarus Luan:	I have ARed quite a few "legal" prims. I have NEVER ARed for a megaprim
[11:44]	Dafydd Laville:	Well that is a problem with almost any sized prim.. just likely to happen more frequently and over more borders with megaprims.
[11:44]	Trent Platthy:	I have, dozens of times.
[11:44]	Jrift Sol:	if the encroachment problem could be solved generally, would megaprims still be controversial, andrew?
[11:45]	Seifert Surface:	its very hard for someone with a 16m^2 plot to do that much harm to a sim, with megaprims you can do it very easily
[11:45]	Scalar Tardis:	forcing an absolute prim / land barrier is not useful because sometimes neighbors are willing to allow encroachment to save prims for both sides.
[11:45]	Talarus Luan:	I don't think they exacerbate the problem any more than legal prims. I think that is more of a theory than a practical issue.
[11:45]	Andrew Linden:	So, since we don't yet have a solution to that problem I'm against raising the max prim sizes until we do.
[11:45]	Dafydd Laville:	What if there were a charge to create (no copy) or perhaps rez prims over a certain size?
[11:45]	Rex Cronon:	they can already do it seinfert
[11:46]	Andrew Linden:	Telarus, only because so many prople now have megaprims in their inventory.
[11:46]	Seifert Surface:	dafydd: that's a whole huge new can of worms
[11:46]	Talarus Luan:	Prims encroach -> ask neighbor to move/remove/resize -> AR if he doesn't -> LL removes. Doesn't matter if it is a megaprim or a tiny prim.
[11:46]	Morgaine Dinova:	Short of a solution to encroachment, why not find a way of reducing the impact it has? Then megaprims don't have to be treated specially.
[11:46]	Talarus Luan:	Exactly. You're not solving the problem by limiting them
[11:46]	Gaius Goodliffe:	/me nods.
[11:46]	Scalar Tardis:	in general I assume more megaprims sets should not be in circulation because that just increases their usefulness and the greater likelihood of this becoming more of a problem sooner than later.
[11:46]	Talarus Luan:	In fact, you make it worse, because the sizes available are not exact-fit, so people may encorach a "little bit"
[11:46]	Seifert Surface:	scalar: how does that follow?
[11:47]	Seifert Surface:	griefing does not need precision
[11:47]	Scalar Tardis:	well the common sets out there now are weird sizes that just don't work in many buildings situations
[11:47]	Gaius Goodliffe:	Seifert: NOT griefing sometimes does.
[11:47]	Scalar Tardis:	so that is a factor against using them vs normal 10m building prims
[11:47]	Seifert Surface:	example?
[11:47]	Dafydd Laville:	The megaprims are out there already. Restricting creation of them will not prevent griefers.
[11:48]	Rex Cronon:	right dafydd
[11:48]	Talarus Luan:	it won't prevent any of the reasons used against them.
[11:48]	Seifert Surface:	scalar: if you're not a griefer (or otherwise clueless about prim encroachment etc), I dont think there is any reason not to allow you to use them
[11:48]	Scalar Tardis:	I am just trying to find out what the consequences are of using new megaprim sets following this latest bug discovery
[11:49]	Morgaine Dinova:	Arbitrary limitations on item size in a virtual world are a pretty wierd concept, actually. Bizarre.
[11:49]	Talarus Luan:	The question, and it is a correct one to ask, is do they really exacerbate the problems? In my experience, it is a LOT easier to spot and report encroachment/griefing with megaprims than with legal prims.
[11:49]	Arawn Spitteler:	/me can see good uses, for prims that contain Access Lines, if the Access Lies don't want a VisibilityOption
[11:49]	Andrew Linden:	I've seen megaprims used effectively, correctly, and well for nice content... all of it on private estates
[11:49]	Creem Pye:	hmm if I wanted to grief with dynamic megaprims, I would create a ball of 255 10x10x10m torii, with each prim right at the link limit edge
[11:49]	Scalar Tardis:	should I spill the beans? hmm.
[11:49]	Andrew Linden:	I'm in favor of megaprims, but I want to fix some other stuff first. So I say "no" to just opening them wide open.
[11:49]	Gaius Goodliffe:	megaprims are used effectively, correctly, and well for nice content everywhere, you don't even notice it most of the time.
[11:49]	Rex Cronon:	u don't need that creem
[11:50]	Talarus Luan:	I use them very well on the mainland, Andrew. So do some of my neighbors.
[11:50]	Dafydd Laville:	I have used some in Houghton.. but it is difficult to work with a limited set.
[11:50]	Seifert Surface:	again, maybe the issue is that nobody wants to spend time coding a stopgap measure which will go obselete when the real solution is there, when they could be coding the real solution?
[11:50]	Talarus Luan:	What stopgap measure needs to be coded?
[11:50]	Seifert Surface:	user interface
[11:50]	Talarus Luan:	You have a policy for ALL prims. Megas are no exception.
[11:50]	Arawn Spitteler:	/me thinks Seifert can say that again, so he'd better be acknowledged:
[11:50]	Andrew Linden:	Enabling megaprims outright wouldn't be much coding.
[11:50]	Rex Cronon:	what is the problem with the gui?
[11:51]	Seifert Surface:	hm
[11:51]	Gaius Goodliffe:	(Indeed, a quick code revert on a certain bug would fix that. ;)
[11:51]	Talarus Luan:	"feature" :P
[11:51]	Seifert Surface:	well, you need gui for stuff like knowing you are overlapping your parcel
[11:51]	Andrew Linden:	It really is the "theory" as Telarus puts it, of causing other issues to suddenly go from "critical" to "blocker".
[11:51]	Talarus Luan:	We already have that
[11:51]	Talarus Luan:	It is called "property lines" and View->Land Owners
[11:51]	Rex Cronon:	no u don't. that is what property lines are for
[11:52]	Andrew Linden:	No, I would put "prim encroachment" at "critical", but I fear that if we gave everybody megaprims on demand it would go "blocker".
[11:52]	Talarus Luan:	I can PRECISELY place even the tiniest prim on my border and not have it encroach
[11:52]	Seifert Surface:	it would be good to have something to tell a newbie that their prim is encroahing
[11:52]	Seifert Surface:	they dont know about parcel boundaries
[11:52]	Rex Cronon:	yes, their neighbour will do that:)
[11:52]	Seifert Surface:	hmm
[11:52]	Talarus Luan:	Andrew: It wouldn't change. Megaprims __already__ __exist__.
[11:52]	Seifert Surface:	youll have sandboxes full of newbie megaprims
[11:53]	Andrew Linden:	Yes Seifert, I was planning on doing that in revision 2 or 3 of the prim encroachment project
[11:53]	Seifert Surface:	that they made with the ui
[11:53]	Arawn Spitteler:	Newbies don't know the value of 117
[11:53]	Seifert Surface:	newbies also dont have access to the existing megaprims
[11:53]	Talarus Luan:	If people want to use megas, they can and they will.
[11:53]	Talarus Luan:	Sure they do
[11:53]	Gaius Goodliffe:	I think it would be fair to say that most people think the solution here (of preventing their creation completely) is worse than the problem. Not to mention not actually solving the problem.
[11:53]	Creem Pye:	"your object "20x20m" has been returned to inventory due to parcel owner return" would teach them quickly, seifert
[11:53]	Andrew Linden:	but haven't been talking much about that these days... someday when we have the basics working we'll start on real-time overlap feedback.
[11:53]	Seifert Surface:	talarus: but the least experienced people dont have megas
[11:53]	Dafydd Laville:	I had a set of megaprims before I new what they were.
[11:53]	Talarus Luan:	"Hey, can I have a box of those?" "Sure! *gives box of megaprims to newbie"
[11:54]	Jrift Sol:	megaprims only exist to those who know about them. if newbies could suddenly stretch a prim to 256m3, there'd be big issues all over the place
[11:54]	Seifert Surface:	creem: yes, but it wont be returned until the code for that is in
[11:54]	Seifert Surface:	jrift: yes, and that's an issue for sandboxes that is independent of parcel boundaries
[11:54]	Dafydd Laville:	WEll when I was a newbie I got megaprims.. I could have easily rezzed them just to see what they were.
[11:54]	Arawn Spitteler:	So, Parcel Owner Return, needs a way to observe object overlap.
[11:54]	Rex Cronon:	if such "big" problem with newbies than don't allow megaprim creation if user less than 1 month old
[11:55]	Gaius Goodliffe:	Jrift Sol -- I agree. But it wouldn't be any more of a problem for them to stretch a prim to 40m as it is for them to make a 40m object using link, which they can do now easily enough.
[11:55]	Talarus Luan:	I don't think that it is that much of a problem.
[11:55]	Andrew Linden:	ok, so to sum things up... (1) Telarus and others think megaprims should be re-enabled pronto.
[11:55]	Seifert Surface:	rex: maybe, although there isnt anything else like that, there are no precedents for that
[11:55]	Dafydd Laville:	/me chuckles
[11:55]	Dafydd Laville:	Moving on? I apologise if this has been jira'd but are many having issues with (not) colliding with things that aren't there?
[11:55]	Jrift Sol:	i think the encroachment problem needs to be resolved first
[11:55]	Scalar Tardis:	well, before the server update finished, I created about a 1000-prim set of the most common small megaprim sizes. I am assuming I should not be giving this out. Best to not annoy andrew. :-)
[11:55]	Gaius Goodliffe:	Re (1) Up to linkset size. It'd be against over that until other problems are solved.
[11:55]	Seifert Surface:	people have argued for such restrictions for self replication functions as well, and people didnt go for that idea
[11:55]	Andrew Linden:	(2) Andrew Linden is not convinced and still thinks some problems should be fixed first.
[11:55]	Talarus Luan:	We already gank newbies in our sandbox at the Isle of Wyrms (well, we did when it was public, anyway) for rezzing huge builds that obscured everything.
[11:56]	Rex Cronon:	there weren't many places like sl before sl existed. were there:)
[11:56]	Talarus Luan:	No megaprims involved.
[11:56]	Creem Pye:	dafydd, like ghost prims?
[11:56]	Seifert Surface:	scalar: why not give it out? how can it possibly make anything worse than it is?
[11:56]	Dafydd Laville:	I made some doors earlier.. and had to make them p[hantom when open so that avatars could walk through the portal.
[11:56]	Andrew Linden:	(3) linkability... yes I'm in favor of increasing it post some other stuff.
[11:56]	Creem Pye:	er ghost objects
[11:56]	Dafydd Laville:	The doors used erm.. I forget the function.. let me find the script
[11:56]	Jrift Sol:	hands up who doesn't own at least one big prim ;)
[11:57]	Dafydd Laville:	ah yes.. llRotLookAt
[11:57]	Seifert Surface:	rex: im talking about sl history
[11:57]	Dafydd Laville:	I know what the objects were.. they were the doors.. that had been rotated so they weren;'t in the way
[11:57]	Arawn Spitteler:	/me was working with a 20X60 the other day, but it wasn't solid across the sim border.
[11:57]	Andrew Linden:	/me has a plan for collisons across region boundaries
[11:58]	Gaius Goodliffe:	yay!
[11:58]	Andrew Linden:	collisions across region boundaries are are pre-requisite for megaprim liberation, in my mind.
[11:58]	Creem Pye:	how about collisions with a dynamic object in an adjacent sim?
[11:58]	Kitto Flora:	I need a 256x40x0.01 :)
[11:58]	Rex Cronon:	there was time in sl when there was no flexi, but now they r here:)
[11:58]	Talarus Luan:	OK. Tell ya what. If the practical worry is that less-experienced users having the ability to harm themselves and others with megaprim creation is the "issue du jour", then fine. Leave the build limit in the production client, but turn it off on the sim. That way, it will require a third-party viewer, which VERY few newbies will use until they are in-the-know enough to understand the implications.
[11:58]	Seifert Surface:	andrew: that id disagree with
[11:58]	Creem Pye:	judging by how region crossings go, that woudl seem pretty difficult to me
[11:58]	Andrew Linden:	More context on that Creem please.
[11:58]	Creem Pye:	if gaius' airsihp is flying in the next sim, but overlaps into my sim,
[11:58]	Andrew Linden:	oh really seifert? Elaborate please.
[11:59]	Creem Pye:	could I stand on it, and interact as though it was in my sim?
[11:59]	Seifert Surface:	sculpties and lltargetomega both give examples of physics not macthing graphics
[11:59]	Creem Pye:	(assuming it's flying around)
[11:59]	Andrew Linden:	Static objects first Creem, since they're easier. Then dynamic.
[11:59]	Seifert Surface:	people deal with it fine
[11:59]	Gaius Goodliffe:	One step at a time. :)
[11:59]	Talarus Luan:	Personally, for me, I just want megaprims for static builds.
[11:59]	Seifert Surface:	if you use a megaprim cross sim, youre building badly
[12:00]	Talarus Luan:	I like them to be phantom OPTIONALLY (floors aren't floors if you fall through them)
[12:00]	Creem Pye:	hehe yeah. personally I support that, because it should make region crossings quicker too, if inter-sim collisions will work
[12:00]	Dafydd Laville:	Likewise though I may want to use up to 40m2 for dynamic.
[12:00]	Seifert Surface:	anyone who has built a road across sims knows about this problem
[12:00]	Rex Cronon:	unless is a bridge or a road seifert
[12:00]	Andrew Linden:	no, I don't like the idea of phantom forced on megaprims either
[12:01]	Scalar Tardis:	actually I see nothing wrong with having megaprim creation being a paid feature of accounts. sort of like the way the something awful forums charge people in order to post. if you mess up, you lose your megaprims rights and must pay to get it back.
[12:01]	Andrew Linden:	although... I'm against most >256 megaprims and think that most of those should mabye be forced phantom
[12:01]	Gaius Goodliffe:	/me nods.
[12:01]	Dafydd Laville:	Agreed
[12:01]	Talarus Luan:	As it stands, the only people who were creating megaprims as it was were the experienced users
[12:01]	Talarus Luan:	I don't have a problem with a 256m limit on megaprims period, for now.
[12:01]	Gaius Goodliffe:	Anything over 256m is doubtless being used for visual effect anyway, since they're already phantom outside their own sim.
[12:01]	Dafydd Laville:	I think that is the point Talarus, if it was opened everyone would be making them.
[12:01]	Seifert Surface:	scalar: requires a new thing for ll customer service
[12:01]	Andrew Linden:	Yes Telarus, that would be a valid solution if it were easier to code up and manage than other solutions.
[12:01]	Jrift Sol:	/agrees with talarus
[12:02]	Kitto Flora:	Assuming encroachment is not an issue, because you have rights in both sims, then it would seem that cross-border penetration of `10M max would be sufficient, as a moving av has eneterd the new region before 10M.
[12:02]	Arawn Spitteler:	/me could think of a valid use, for slightly greater than 256
[12:02]	Talarus Luan:	Well, there's no code involved. If the production client can't make them, then newbies won't hurt themselves and others
[12:02]	Andrew Linden:	yes Gaius, everything >256 is being used for valid reasons... except those that are not, or are just stuipid building techniques
[12:02]	Gaius Goodliffe:	Kitto: I think I've gone past the 10m mark before...
[12:02]	Scalar Tardis:	a skybox prim would definitely be useful, which would eliminate the need for >256m prims since they are mostly used for skyboxes anyway
[12:03]	Andrew Linden:	it is hard to program enough intelligence into the simulator to know the good from the bad with high accuracy
[12:03]	Talarus Luan:	Enabled on the sim, NOT enabled in the production client.
[12:03]	Creem Pye:	/me has probably gone past the 100m mark before, when crossing sims
[12:03]	Rex Cronon:	wow. since when is building a building over 256m, is "stupid"?
[12:03]	Gaius Goodliffe:	Might I add, it would be *really* nice if we could exclude the huge prims, or selectively exclude any selected prims, from the minimap.
[12:04]	Yuu Nakamichi:	or set the height of the cut off on the minimap?
[12:04]	Seifert Surface:	re: there are dumb ways to use a 256m prim
[12:04]	Seifert Surface:	rex that is
[12:04]	Creem Pye:	yeah, it's really obnoxious to have some horizon prim blocking everything else on the minimap :-|
[12:04]	Andrew Linden:	Rex, I was referring to the possibility of putting a >256 prim horizontal and overlapping a bunch of neighboring regions.
[12:04]	Scalar Tardis:	yes, that is a question mark. limit 256x256 on the X Y axis but not necessarily on the Z. >256m is practical for a tall building sim
[12:04]	Dafydd Laville:	That solution, Talarus would need to be accompanied by making megaprims no transfer only.
[12:04]	Gaius Goodliffe:	Oh yeah. I think it's kinda silly that the minimap shows me prims at 1000m+ while I'm on the ground.
[12:05]	Talarus Luan:	It's not stupid, Rex. :) But a big building can still be made of multiple megas. Though viewer draw distance is gonna wreak havoc with it
[12:05]	Jrift Sol:	again, it's back to encroachment. that does seem to be the underlying problem
[12:05]	Andrew Linden:	well... just rotate it Scalar and you've got the same problem... local Z along global XY
[12:05]	Rex Cronon:	ok. that sounds better, but there are times when u want/need to build something over 256
[12:05]	Creem Pye:	using 16 megaprims for a building to reach 4096m tall doesn't seem too horrible to me
[12:05]	Seifert Surface:	you dont want a prim to go illegal just by rotating it
[12:06]	Creem Pye:	(horrible in terms in prims needed)
[12:06]	Scalar Tardis:	heh, I guess so Creem.
[12:06]	Andrew Linden:	alright, well it is good to know the megaprim issue is nice and hot. Good input for adjusting the priority on the relevant projects.
[12:06]	Talarus Luan:	Anyway. I'm just arguing in circles now. My basic premise is that removing megaprim creation is a signficant net negative for SL.
[12:06]	Seifert Surface:	heh, yes
[12:06]	Gaius Goodliffe:	When it comes to input, we're always full of it. :)
[12:06]	Andrew Linden:	Thank you all for coming.
[12:06]	Andrew Linden:	As I said...
[12:07]	Dafydd Laville:	ON the issue of being lbocked by stuff that isn't there?
[12:07]	Jrift Sol:	thank you andrew
[12:07]	Dafydd Laville:	blocked*
[12:07]	Scalar Tardis:	Hey, um, talarus, would you like to buy a 1000 megaprim set? ;)
[12:07]	Dafydd Laville:	Thank you, Andrew.
[12:07]	Arawn Spitteler:	/me had an idea, for the border problem, but it means sending the whole Physics Engine back to Drafting: What would be the cost benefit, of moving the Physics Engine Client Side?
[12:07]	Rex Cronon:	talarus did u create bug on jira?
[12:07]	Talarus Luan:	Yes, thank you for listening, Andrew.
[12:07]	Andrew Linden:	LL will be having a meeting on Thursday about megaprims, so I should have more info about the state of LL's consensus on Thursday's hours.
[12:07]	Gaius Goodliffe:	Psst, don't buy it. Look behind you. :)
[12:07]	Talarus Luan:	I have some, Scalar ;)
[12:07]	Kitto Flora:	Byebye Nadrew et al.
[12:07]	Yuu Nakamichi:	I hope to hear about the TD issue as well
[12:07]	Kitto Flora:	Lunchtime!
[12:08]	Trent Platthy:	So are we going back to talking about H4 now?
[12:08]	Yuu Nakamichi:	(Sidewinder's project)
[12:08]	Gaius Goodliffe:	Havok4? What's that? ;)
[12:08]	Andrew Linden:	Moving the physics engine completely client side would be... a lot of work and would open up all sorts of problems.
[12:08]	Arawn Spitteler:	/me couldn't buy, as it too 7,000 too long to load.
[12:08]	Andrew Linden:	H4 is old news.
[12:08]	Andrew Linden:	And most of the bugs are fixed.
[12:08]	Andrew Linden:	Most of them.
[12:08]	Rex Cronon:	h4 office hour is from 11 to 12, trent
[12:09]	Jrift Sol:	h4 works great, it's just a bit of tuning, as far as i can tell
[12:09]	Andrew Linden:	Hehe, my chats are arriving in reverse order.
[12:09]	Trent Platthy:	We spent a good deal of it talking about megaprims, but whatever...
[12:09]	Talarus Luan:	Sorry. :)
[12:09]	Trent Platthy:	I'll just show up for the megaprim discussion and start talking about my favorite scripting bugs.
[12:09]	Andrew Linden:	Trent, you had other topics to bring up?
[12:09]	Talarus Luan:	I had to get the kink out of my tail over it. ;)
[12:09]	Trent Platthy:	No, I like my plan better, thanks.
[12:09]	Andrew Linden:	Script bugs? Or Havok4 bugs?
[12:10]	Trent Platthy:	Discussing scripting bugs at the megaprim meeting.
[12:10]	Andrew Linden:	ok sounds like a good plan.
[12:10]	Dafydd Laville:	I believe my issue is a H$ bug.
[12:10]	Dafydd Laville:	I could go get the doors to demonstrate.
[12:10]	Andrew Linden:	Is that really James Miller?
[12:10]	Gaius Goodliffe:	H$? The currency of physics?
[12:10]	Creem Pye:	keep in mind that sometimes the Windlight viewers are lazy to draw object updates, dafydd
[12:10]	Scalar Tardis:	thanks gaius, you made me feel less guilty about making my 1000.
[12:10]	James Miller:	Yes, yes it is. :)
[12:11]	Gaius Goodliffe:	lol yw
[12:11]	Dafydd Laville:	ok so I can't type :p
[12:11]	Andrew Linden:	It's been a long time James.
[12:11]	James Miller:	It has, Andrew. How have you been? :)
[12:11]	Creem Pye:	it's pretty common for moving child prims to not be drawn in their new position, in my experience =\
[12:11]	Andrew Linden:	Alright... if it weren't for megaprims ;-)
[12:11]	James Miller:	hah.
[12:11]	Dafydd Laville:	I've seen that frequently.. but click on the child will update it
[12:11]	Talarus Luan:	My notes I was operating from :)
[12:12]	Talarus Luan:	In case you want to pass them out at your meeting on Thursday
[12:12]	James Miller:	I saw you had office hours today, I figured I'd come and stalk you for a minute.
[12:12]	Andrew Linden:	Well you arrived a bit late James.
[12:12]	James Miller:	Aw, shucks.
[12:13]	James Miller:	I'm not good with time zones.
[12:13]	Andrew Linden:	Show up on time and I try to answer your questions ;-)
[12:13]	Yuu Nakamichi:	ao on
[12:13]	Andrew Linden:	Cheers everyone... gotta go.
[12:13]	Talarus Luan:	/me points to the clock in the top-right of the screen.
[12:13]	Arawn Spitteler:	SLT is a clock in San Francisco
[12:13]	James Miller:	Hah.
[12:13]	Andrew Linden:	/me grabs the transcript...
[12:13]	Dafydd Laville:	Thanks, Andrew.
[12:13]	James Miller:	And poof, he's gone.
[12:13]	Jrift Sol:	thanks again andrew, see you next time
[12:13]	Rex Cronon:	bye andrew
[12:13]	Talarus Luan:	Thanks for putting up with me, folks. ;)
[12:14]	Jrift Sol:	lol. he ran
[12:14]	Gaius Goodliffe:	Happy to have you. :)
[12:14]	Arawn Spitteler:	What would it take, to move the Physics Engine Client Side? At least we'd collide with anything we could see.
[12:14]	Creem Pye:	is andrew in SF?
[12:14]	Arawn Spitteler:	Washington State, I think, but SLT is in SF
[12:14]	Rex Cronon:	would we a a quadcore arawn:)
[12:14]	Talarus Luan:	In general, the client doesn't often have enough of the dataset to do proper simulation.
[12:14]	Creem Pye:	ah, but PST anyway. guess he was in a hurry to eat =P
[12:15]	Jrift Sol:	it's got to be server side for object->object interactions. it'd be inconsistent between viewers otherwise