Simulator User Group/Transcripts/2011.06.24

From Second Life Wiki
Jump to: navigation, search

Simulator_User_Group

Prev 2011.06.21 Next 2011.06.28

List of Speakers

Andrew Linden
Ashiri SandsSimulator_User_Group
Prev 2011.06.21 Next 2011.06.28
Kallista Destiny Latif Khalifa
Michelle Resistance NielArcher
Odessa Ibanez Pauline Darkfury
Peewee Musytari Simon Linden
WiZZKiD

Transcript

[16:09] Jonathan Yap: ?

[16:09] Simon Linden: ... and Andrew, who's here :)

[16:09] Kallista Destiny: Hi Andrew

[16:09] Andrew Linden: Hello. Sorry I'm late.

[16:10] Andrew Linden: I forgot (and was busy filing some minor bug)

[16:10] Simon Linden: We were just going over the release notes, I mentioned the LSL changes in BlueSteel

[16:10] Pauline Darkfury: Hi Kelly, Andrew

[16:10] Kallista Destiny: Simon was filling us in. on the Greatest and latest

[16:10] Andrew Linden: Ok carry on. That's all news to me

[16:10] Jonathan Yap: Is Kellly an invisible avatar?

[16:10] Simon Linden: LeTigre and Magnum kept the mono and mesh code

[16:10] Pauline Darkfury: oh, lol, just me misreading chat

[16:11] Pauline Darkfury: I thought Kelly had arrived, but it was "hail Kelly", not "Hai Kelly" ;)

[16:11] Simon Linden: That's pretty much it for releases ... Andrew, did you have other news?

[16:11] Andrew Linden: A little bit of news...

[16:12] Andrew Linden: I finished the 8-neighbor code, and also cross-region encroachment returnability stuff.

[16:12] Michelle Resistance: :o

[16:12] Odessa Ibanez: yayyyyyyy!!!!

[16:12] Kallista Destiny: Woot!!!!!

[16:13] Ashiri Sands: cool

[16:13] Andrew Linden: The 8-neighbor code will be disabled for a long time, until it can be supported everywhere (we'll leave the world in 4-neighbor mode until then)

[16:13] Andrew Linden: at which point we'll turn on 8-neighbor mode as a config change all at once

[16:13] Andrew Linden: the 8-neighbor stuff is just to support cross region encroachment return across corners really

[16:13] Kallista Destiny: Please tll me so I can wear my Grid crash box...

[16:14] Andrew Linden: Note: priviate estates that only have corner neighbors will NOT be able to see their neighbors in 8-neighbor mode

[16:14] Michelle Resistance: Okay, good, I was just about to ask. ^_^

[16:14] Andrew Linden: the neighbors only show up if there is a regular (4-neighbor) region to bridge

[16:14] Andrew Linden: Also, I had to fix a few minor "encroachment return" bugs today.

[16:15] Andrew Linden: such as, the amount of overlap before an objedct is determined to be encroaching was incorrect on mesh simulators

[16:16] Andrew Linden: this cross-region encroachment work only exists in the mesh-prep project (but not yet deployed to Magnum)

[16:16] NielArcher: how much overlap triggers return?

[16:16] Andrew Linden: So... I will now (Monday) start talking to the people who would decide when/how to enable encroachment on the mainland

[16:16] Kallista Destiny: how is encroachmet determined? That is is it an absoulte distance or a % of the prim?

[16:16] Andrew Linden: maybe we can start testing/enabling it when this hits an RC channel -- we'll see.

[16:17] Andrew Linden: Right now a box object can be butting right up to the prim (flush to the edge) and it is NOT considered to be overlapping

[16:17] Kallista Destiny: nods

[16:17] Andrew Linden: but if you move it 1 cm over then it is returnable.

[16:17] Andrew Linden: Kallista, encroachment is determined using the physics engine

[16:18] NielArcher: excellent

[16:18] Michelle Resistance: ^

[16:18] Andrew Linden: we build a "collision proxy" for the selected object

[16:18] Kallista Destiny: nods so basically ANY encroachment

[16:18] Latif Khalifa: so sculpts will not work? ;)

[16:18] Andrew Linden: and then make a collision query between that proxy and a bunch of shapes already built for the parcels (in a separate "Physics World" instance)

[16:19] Andrew Linden: Sculpts definitely don't collide according to their visual shape

[16:19] Pauline Darkfury: presumably phantom and new mesh no-physics stuff will work ok?

[16:19] Kallista Destiny: This is done when a return attempt is made?

[16:19] Andrew Linden: and might not even collide according to their collision shape -- they may collide as their box -- I'll have to test.

[16:19] Andrew Linden: Mesh objects (when they are enabled) will collide for encroachment as their bounding box

[16:20] Andrew Linden: this is because a mesh object will be able to collide very differently from their visual representation

[16:20] Michelle Resistance: Will this have any effect on the physical aspects of prims lying across region borders?

[16:20] Andrew Linden: and the simulator doesn't have their visual collision info

[16:20] Latif Khalifa: I was wondering why proper sculpt supprot was not added now that alot of stuff was added to the physics

[16:20] Andrew Linden: so just to be safe, we use the boundin box

[16:20] Michelle Resistance: Is a convex hull not possible?

[16:21] Andrew Linden: Michelle, no objects will not collide with other objects across region boundaries. That is a separate project that we want to work on (maybe even soon).

[16:21] Michelle Resistance: Okay.

[16:21] Andrew Linden: Mesh objects that use convex hulls for collisions could also use a different "visual" shape

[16:22] Andrew Linden: so they still collide like boxes -- same problem

[16:22] WiZZKiD: hey

[16:22] Andrew Linden: er... collide like boxes for encroachmen tests

[16:22] Andrew Linden: WiZZKiD, welcome.

[16:23] Andrew Linden: Pauline, yes phantom, volume-detect, and no-shape objects should collide properly for encroachment tests

[16:23] Andrew Linden: that is one of the reasons why we build a "proxy shape" when doing the encroachment collision queries --

[16:23] WiZZKiD: hey

[16:23] Andrew Linden: in general the actual shapes we have for normal collisions are not the ones we want to use for encroachment tests

[16:24] WiZZKiD: tommy davis

[16:24] Andrew Linden: That is all the news I have.

[16:25] Andrew Linden: I think I'll be working to make sure the mesh project is not blocked -- over the next few weeks.

[16:25] Pauline Darkfury: Thanks for the update, will be very glad to have that stuff when it arrives, particularly on mainland

[16:25] Andrew Linden: So I'll be fixing any bugs or blockers there as my main priority.

[16:25] NielArcher: same here, will be great to see it in action

[16:25] Andrew Linden: Dunno if I'll have much time for general maintenance. We'll see.

[16:25] Kallista Destiny: Yeah I'm sure that mainlanders whill be very happy

[16:25] Pauline Darkfury: One "fix pending" bug I'd like to raise - SVC-6905

[16:25] JIRA-helper: http://jira.secondlife.com/browse/SVC-6905

[#SVC-6905] Immortal prims: Objects resisting take, delete, return, and autoreturn

[16:26] Andrew Linden: That is fixed in one of the RC channels I belive.

[16:26] Michelle Resistance: Is that related to ghost prims?

[16:26] Pauline Darkfury: Any idea when we might see the fix get deployed?

[16:26] Jonathan Yap: Is there a maint release scheduled for any time soon?

[16:26] Pauline Darkfury: Oh, cool

[16:26] Simon Linden: That fix is in the release pipeline, I believe

[16:26] Andrew Linden: Simon fixed that buig.

[16:26] Andrew Linden: Not in RC yet Simon?

[16:26] Simon Linden: yeah, let me follow the paperwork and see if I can find out where it is....

[16:26] Kallista Destiny: Immoral prims should be banned to Zindra.

[16:27] Pauline Darkfury: Yup, just wanted to ask, as it's still happing fairly often, and a bit of a pain

[16:27] Andrew Linden: I thought it was in FireEngine-18 (maint-server project) which went out on RC this week.

[16:27] Simon Linden: Right, that should be on Blue Steel now

[16:28] Pauline Darkfury: Excellent, thanks :)

[16:28] NielArcher: it's listed in BlueSteels notes

[16:28] Pauline Darkfury: Ok, I'm behind on release note reading this week

[16:29] Andrew Linden: Michelle, I don't think that was related to "ghost prims". It was a distinct bug where an object's "delete" status would get stuck due to a lost packet.

[16:29] Kallista Destiny: Is that because the Magnum notes are so extensive

[16:29] Michelle Resistance: Ah, okay.

[16:29] Andrew Linden: And "ghost prims" is a bit ambiguous. We've had ghost colliders, collisionless (viewer-side) ghosts, and others.

[16:30] Pauline Darkfury: There was some unofficial speculation that it could have been the latest form of what once were ghost prims, but that was never confirmed

[16:30] Simon Linden: it's possible ... these were essentially stuck in limbo as they were being removed from the region

[16:30] Andrew Linden: No the "immortal prims" bug was introduced by our "object delete cleanup" project which was in one of the main-servers several weeks ago.

[16:30] Meeter: Timecheck : User Group is half over

[16:31] Simon Linden: The sim was waiting for a packet to arrive, but had already missed it

[16:31] Kallista Destiny: Does th internal communication us UDP mostly?

[16:31] Simon Linden: so it would wait ... and wait .. and wait, and not let you deleted it since it thought it was already doing that

[16:31] Michelle Resistance: I was about to ask that. XD

[16:31] Pauline Darkfury: Yup, makes sense, Simon

[16:32] Simon Linden: it was basically a hole in the error checking

[16:32] Simon Linden: Now there's a timeout and it recovers much more gracefully

[16:32] Andrew Linden: Most of the very old internal communication uses UDP. As we add new stuff we usually use different protocols -- it is hard or frowned-upon to add new UDP messages.

[16:33] Kallista Destiny: 'Depreciated'

[16:33] Simon Linden: yeah, there is very little of the internal traffic that shouild be on UDP ... eventually

[16:33] Pauline Darkfury: Well, I'd hope that UDP is still acceptable for classes of stuff where loss isn't important. It does have useful purposes sometimes

[16:33] Andrew Linden: For instance, the "get neighbor region's parcel info" that I added uses a "push via TCP" system.

[16:34] Simon Linden: and if you're following the viewer code, we're trying to replace UDP messages to the viewer with tcp caps whenever possible

[16:34] NielArcher: Jane, try the other "'empty' chair

[16:34] Andrew Linden: Whats' that old SVC bug about crossing regions on vehicles into parcels where you're not allowed?

[16:35] Simon Linden: I think someone's in that other one, just a bit under the table today

[16:35] Latif Khalifa: svc-22

[16:35] JIRA-helper: http://jira.secondlife.com/browse/SVC-22

[#SVC-22] Vehicles crossing region borders aren't always treated as vehicles and can get incorrectly returned if the destination parcel is no-entry or parcel-full

[16:35] Kallista Destiny: I presume you added an accellerated 'close this connection' to keep the close wait2 states from hanging around?

[16:35] Andrew Linden: Yeah -22, that's the one.

[16:35] Andrew Linden: I think I could actually fix that one now.

[16:35] Andrew Linden: I'll try to look into it next week.

[16:35] Latif Khalifa: hehe

[16:35] Simon Linden: ... we also have another engineer who's looking at region crossings now

[16:35] Kallista Destiny: You'll make lots of drivers /polits happ

[16:36] Michelle Resistance: How many low-numbered SVC issues are still around?

[16:36] Simon Linden: and was horrified as usual when she learned how vehicles and their riders cross

[16:36] Kallista Destiny: piolts

[16:36] Andrew Linden: BTW, I started to look into how to throttle the llDialog() stuff, but got sidtracked. It should be too hard. I'll try again next week.

[16:38] Andrew Linden: Oh yeah I was going to mention... I'm looking for someone who might want to give some advice to a small university that is thinking about opening up a space in SL.

[16:39] Michelle Resistance: {{ I have a suggestion for addressing vehicle region crossings, actually. Should I tell you guys, or is that better directed at the JIRA or other engineer? }}

[16:39] Andrew Linden: A friend of a friend was contacting me and while I'm good at talking about the technical aspects of SL, I'm not an expert at how it is actually used.

[16:39] Andrew Linden: Go ahead Michelle.

[16:40] Michelle Resistance: Well, one of the biggest issues is that a vehicle or other object doesn't have many options for *not* crossing a region border.

[16:40] Michelle Resistance: You can set the DIE_AT_EDGE flag, or try to actively avoid it, but...

[16:40] Simon Linden: hmm, that's a good point ... BOUNCE_AT_EDGE might be useful

[16:40] Andrew Linden: You want to be able to specify the object's behavior when hitting a region boundary?

[16:40] Michelle Resistance: Yes, exactly.

[16:40] Michelle Resistance: BOUNCE_AT_EDGE would be amazing.

[16:41] Pauline Darkfury: Yeah, that seems like a potentially very useful addition

[16:41] Andrew Linden: Hrm... it would be easy to add.

[16:41] Pauline Darkfury: Maybe even a collision_edge companion event?

[16:41] Simon Linden: We also could do a better job at querying the next region to make sure it's OK for an object to enter

[16:41] Michelle Resistance: And, actually, there are useful cases for bouncing at sim edge and bouncing at world edge.

[16:41] Jonathan Yap: Wrap around would be cool, too, go out at one edge and appear at the corresponding other side (e.g. depart at bottom, arrive at top)

[16:41] Andrew Linden: We already have logic for figuring out if a bounce is necessary

[16:41] Jane1 Bookmite: Andrew, I teach at university, technology, in sl help manage a sim and host. If you think I could, I'd be happy to help

[16:41] Andrew Linden: when hitting an edge the doesn't have a neighbor, for example

[16:41] Simon Linden: haha, taht would be great on combat sims, Jonathan

[16:41] Andrew Linden: so that code could pivot on another bit

[16:42] Jonathan Yap: yes, give you much more apparent space

[16:42] Pauline Darkfury: Oh, and if considering collision_edge as an event, knowing if you've impacted on no-object-entry parcel boundaries would be good

[16:42] Michelle Resistance: Well, if you do create something like that, please keep in mind the distinction of sim edge and world edge.

[16:42] Andrew Linden: Jane1, I'm also open to advice. Maybe I should just refer them to the SLED list, or maybe I should solicit on that list.

[16:42] Pauline Darkfury agrees with Michelle

[16:43] Jonathan Yap: Andrew, put your question out to the AWG group (are you in that one?)

[16:43] Pauline Darkfury: Speaking to other .edu places that have made good use of SL is probably the best for them, tbh

[16:43] Michelle Resistance: I've been working on some physical missiles, which currently have to either run DIE_AT_EDGE or actively flee the edge of the world, which doesn't work very well.

[16:43] Kallista Destiny: Simon, that would be madness in combat sims, believe me, you'd alwasy shoot yourself if you missed your target.

[16:43] Pauline Darkfury: many years since I was in academia, but there always seemed to be good tech interaction between institutions, sharing the infrastructure knowledge, etc

[16:43] Michelle Resistance: Wrapping around is a bit nuts.

[16:43] Kallista Destiny: (assumin the tht wrap retaind direction

[16:44] Michelle Resistance: But an event that would let a script decide the behavior would be useful.

[16:44] Jonathan Yap: Wrap around could be interesting. You'd only shoot yourself it your shot straight s, n, e, w and nothing was in the way

[16:44] Ruth Meeroo whispers: Simon Linden will you please pick me up and snuggle me?

[16:44] Pauline Darkfury giggles at Simon's pets

[16:44] Kallista Destiny: actually ni, although the projectal might have to cross 4 boundries to come bach to the orgin

[16:44] Jonathan Yap: I've played a game with bounce and wraparound, and sometimes both were on at once

[16:44] Simon Linden: these things are getting popular, i wanted to check them out ... it's a pretty impressive system

[16:45] Pauline Darkfury: It's great to see you guys trying stuff like that out, Simon :)

[16:45] Kallista Destiny: or even more if is's shallowly angled

[16:46] Jonathan Yap: You'd have to adjust your mind to the new physics of the world

[16:46] Andrew Linden: So, how would the edge behavior be set? llSetStatus()? or something else?

[16:46] Jonathan Yap: I suppose you could have similar rules for the z direction

[16:46] Michelle Resistance: llSetStatus would be nice.

[16:46] Michelle Resistance: But, just to be clear, we mean physical bouncing, not returning to inventory, correct?

[16:46] Pauline Darkfury: Other thing vaguely related to BOUNCE_AT_EDGE, etc. There's a Jira for it, but I forget the number - a way to prevent scripted objects entering no-script land

[16:46] Andrew Linden: Hrm... actually llSetStatus() params just take TRUE and FALSE, right?

[16:47] Pauline Darkfury: yeah, it's a bitfield, so has to be binary options there

[16:47] Jonathan Yap: A bounce would be a reflection at a region edge

[16:47] Simon Linden: Pauline - that's the kind of thing I was thinking about earlier. The region could check with the neighbor earlier to see if it's OK to send an object over or not

[16:47] Kallista Destiny: on condition and signal condition would be useful... (as in PL\1)

[16:47] Simon Linden: Now it's more of a send-and-see approach

[16:48] Andrew Linden: If we wanted to add a new edge behavior (DIE for example) then the bitfield might not be the best mechanism

[16:48] Pauline Darkfury: Yup, indeed, Simon. I'm thinking of both region-crossing and parcel-crossing cases, to help reduce junk prims that get jammed into no-object-entry & no-script type settings in all situations

[16:48] Kallista Destiny: on_edge( Integer) event

[16:48] Michelle Resistance: STATUS_RETURN_AT_EDGE and STATUS_DIE_AT_EDGE both refer to the edge of the world in the documentation...

[16:49] Pauline Darkfury: Yup, those are edge-of-world, not region boundary, according to docs

[16:49] Andrew Linden: Ah, those already exist.

[16:49] Michelle Resistance: But, physically bouncing off the world edge like an avatar? That'd be awesome.

[16:49] Andrew Linden: Yes I see. Those refer to hitting the void

[16:49] Kallista Destiny: where the integer has useful informarion no entry, no script etc

[16:49] Michelle Resistance: Actually, we wouldn't even need a new event!

[16:49] Andrew Linden: Bounce at the voide boundary is the default

[16:49] Michelle Resistance: Could use collision

[16:49] Simon Linden: This would probably be a per-edge setting too, which is something new

[16:49] Michelle Resistance: Or even land_collision

[16:50] Michelle Resistance: land_collision would be perfect, actually.

[16:50] Simon Linden: I can imagine having different behavior for North, South, East and West

[16:50] Michelle Resistance nods.

[16:50] Pauline Darkfury: makes sense, I think

[16:50] Andrew Linden: Why would we have different behavior at different edges

[16:50] Pauline Darkfury: diagonal cases, if Andrew's diagonal crossings stuff comes to be?

[16:50] Simon Linden: It seems like an interesting idea ... the next step might be to write up a jira with a proposed API

[16:51] Michelle Resistance: But, just like llVolumeDetect modifies the behavior of collision_start and collision_end, the land_collision_start behavior could be modified.

[16:51] Michelle Resistance nods.

[16:51] Andrew Linden: and note, when 8-neighbor paradigm hits there will be a small window for tunnelling through to the corners

[16:51] Simon Linden: well, all crossings now are in the 4 major directions. You do two if you want to go diagonally

[16:51] Michelle Resistance: Which is a bit frustrating.

[16:52] Simon Linden: yeah, I can see where just jumping diagonally makes sense when you're really right at the corner

[16:52] Simon Linden: but then I can imagine people building walls right to the edge and arguing that you can't get through

[16:52] Jonathan Yap: I was doing that for the group chat test, or trying to, but kept getting offset to one side

[16:52] Andrew Linden: the new code will send avatars or objects directly to the corner neighbor if that's where you're supposed to go

[16:53] Simon Linden: nice!

[16:53] Andrew Linden: someone said the "bounce at region boundary" may have already been put into a jira issue

[16:53] Andrew Linden: I'll make a quick search

[16:54] Michelle Resistance: I'm eager to participate in a JIRA discussion for a proposed API. ^_^

[16:55] Meeter: Timecheck : User Group is almost over

[16:55] Jonathan Yap: You could have a sound effect for the bounce event, too "clang!"

[16:56] Andrew Linden: I didn't find it.

[16:56] Michelle Resistance: I don't think we should have a new event for it.

[16:56] Simon Linden: yeah, let's keep it simple if possible

[16:56] Andrew Linden: Hrm... hunting through the bug list is hazardous. I see a bunch of stuff I should work on.

[16:56] Pauline Darkfury: Yup, I just suggested a new event, as I wasn't sure if overloading existing events could have a negative aspect

[16:57] Michelle Resistance: "Hazardous" is such a negative word. =p

[16:57] Kallista Destiny: A new event is less likly to break existing c0ntent

[16:57] Michelle Resistance: Well, it wouldn't really break existing content, because the behavior would only be modified when setting the new status.

[16:57] Simon Linden: There's some resistance to freely adding new events, but that's a long story

[16:57] Michelle Resistance: Just like llVolumeDetect

[16:58] Pauline Darkfury: yup, indeed, Kallista, or make it easier to have fast light code that doesn't have to test lots of alternate cases to figure out what actually happened

[16:58] Andrew Linden: What about objects that bounce "unexpectedly" at region boundaries.

[16:58] Michelle Resistance: How would that happen?

[16:58] Andrew Linden: I could imagine that the feature would be reported as a bug when things bounce when they are not expected.

[16:59] Andrew Linden: That is... how does one "know" that the bouncing at the border is "working as designed"?

[16:59] Michelle Resistance: If someone has to call a function to enable the bouncing, I don't see how that's an issue.

[16:59] Michelle Resistance: Well, I suppose from a scripting perspective.

[16:59] Kallista Destiny: exactly like haveing multiple I/O interrupt vectors, vs 1 I/O interrupt vectors (long skip chains testing for avarious events.

[16:59] Jonathan Yap: Someone could be using some object not knowing that is has bounce/reflect enabled

[16:59] Pauline Darkfury: Yeah, if the new behaviour has to be enabled somehow, existing content should be fine

[16:59] Ashiri Sands: But the end user might think it was a bug

[17:00] NielArcher: because if it's not as designed you get through 2-3 sims before bouncing back when it's just lag ;-)

[17:00] Andrew Linden: yeah, that is what I mean

[17:00] Michelle Resistance: That's more an issue for the content creator using the feature, I think.

[17:00] Simon Linden: it may be a strange experience, depending on the objects involved and what the region looks like

[17:00] Latif Khalifa: I have to take off, have a great weekend everyone.

[17:00] Andrew Linden: not that it is a blocker but I could imagine seeing bug reports: this object of mine cannot cross region boundaries

[17:00] Simon Linden: Take it easy, Latif

[17:00] Pauline Darkfury: Have a good one, Latif

[17:00] Andrew Linden: tc

[17:00] NielArcher: bye Latif

[17:01] Kallista Destiny: well once again a very interesting and useful meeting. Thank you again.

[17:01] Jonathan Yap: Could something akin to ban lines be put up?

[17:01] Pauline Darkfury: People will get used to that as a possible behaviour

[17:01] Simon Linden: They could, Jonathan, but people seem to hate those already

[17:01] Michelle Resistance: I agree with Pauline.

[17:01] Jonathan Yap: you would only see the lines if you are very close

[17:02] Pauline Darkfury: Yeah, ban lines are much hated and banned in places

[17:02] Jonathan Yap: you'd have to be sitting on some object that would be using the bounce setting for them to show

[17:02] Simon Linden: I agree, I don't think this is something that would kill the feature ... it's just an oddity to consider and experiment with

[17:02] Pauline Darkfury: The issues with ban lines are they are both kinda ugly and you don't see them until too late. Can't really solve that, as the solution to one aspect makes the other worse

[17:02] NielArcher: hehe cute Kallista

[17:02] Kallista Destiny: what's wrong with ban lines?

[17:03] Jonathan Yap: I found the other day there are also pass lines, which probably no one has ever seen

[17:03] NielArcher: shame you can't get something like that for FL

[17:03] Michelle Resistance: I've seen them. They're weird.

[17:03] Pauline Darkfury: Spend enough time on mainland, you see pass lines around

[17:03] NielArcher: the Buy a pass lines?

[17:03] Kallista Destiny: like these

[17:03] Kallista Destiny: ?

[17:03] Simon Linden: I was wondering if anyone ever used teh pass system

[17:04] Ashiri Sands: yes

[17:04] NielArcher: yes, haven't seen many of them for over 5 years now

[17:04] Pauline Darkfury: Can't say I've ever seen anyone actually really using them (i.e. people inside who have bought a pass)

[17:04] Ashiri Sands: But only twice have i seen them used

[17:04] Kallista Destiny: There are a few places, but pretty rare I must say

[17:05] Andrew Linden: Yeah, that was a feature idea that was never finished to a point where it could become popular/useful.

[17:06] Michelle Resistance: It doesn't really serve the intended purpose, I think. A pass line doesn't stop you from camming into a museum parcel, for example.

[17:06] NielArcher: I think it was more of a shock to earlier resis. They expected you to pay them to come in, not be charged ;-)

[17:06] Simon Linden: ah, good point ... not yet :)

[17:06] Pauline Darkfury: lol @ Niel

[17:06] Michelle Resistance: Restricting the camera isn't really a good thing, though. :<

[17:06] Jonathan Yap: Is there going to be a writeup for the new server features: invisible avatars and new parcel sound controls?

[17:07] Simon Linden: We'll get something onto the wiki, hopefully next week

[17:07] NielArcher: how do "invisible avatars" differ from what I am now?

[17:07] Ashiri Sands: no slippers?

[17:07] NielArcher: lol

[17:07] Simon Linden: ok, here's my speil on the new features coming out:

[17:07] NielArcher: I mean the bits you can't see ;-)

[17:07] Jonathan Yap: You can set your parcel to invisible avatar mode, then I cannot see you

[17:07] Simon Linden: First, there are now settings to control sounds coming from avatars - gestures, played from the viewer, or from scripted attachments

[17:08] Pauline Darkfury: One question on the invisible avatars. If you can hear them in chat, will you always be able to see them?

[17:08] Simon Linden: You can set 'avatar sounds' to be allowed for everybody, the parcel's group's members, or nobody, just like other options such as rezzing objects or running scripts

[17:08] Simon Linden: Right now, all parcels let everybody make sounds, so that will be the behavior as it rolls out

[17:08] Michelle Resistance: Hah, okay.

[17:08] Simon Linden: They'll be a viewer update that has the new checkboxes soon ... I'm not sure of the date, but the code will be public early next week

[17:08] Simon Linden: This should be useful at concerts, presentations and such where you don't want annoying sounds distracting others

[17:08] Michelle Resistance: { Did the TPV feature of allowing users to mute gestures make it into the mainline client? }

[17:09] Simon Linden: I'm not sure about that, Michelle

[17:09] Simon Linden: The other parcel feature is 'hidden avatars'. If this is on, you only see, and are seen by, avatars that are on the same parcel as you are.

[17:09] Jonathan Yap: Michelle, it was the other way around, that feature started out in the mainline client and was picked up by the tpv

[17:09] Michelle Resistance: Oh, okay.

[17:09] Simon Linden: If it was running here, we'd all see each other normally, but as soon as someone flew off the edge of this parcel, they'd disappear.

[17:09] NielArcher: that's coming along, Jonathan is doing it ;-)

[17:09] Simon Linden: They would no longer see anyone here.

[17:09] Simon Linden: Text chat on channel 0 obeys the same rule ... it's restricted to the parcel

[17:09] Michelle Resistance: That's great for privacy.

[17:09] Simon Linden: The map, mini-map, LSL sensors and script functions all behave normally and can show or interact with all avatars

[17:10] Michelle Resistance: Wait, does that apply to objects, too?

[17:10] Simon Linden: llGetParcelDetails() supports a new PARCEL_DETAILS_HIDDEN_AVS value to read the setting

[17:10] Simon Linden: no, just avatars

[17:10] Michelle Resistance: Okay.

[17:10] NielArcher: I'm wondering about objects too

[17:10] Simon Linden: There's also a special case for vehicles: if they have an avatar on them, the entire vehicle is not visible. This gives us some better performance, plus driver-less vehicles seemed a bit strange when we tried it out.

[17:10] Pauline Darkfury: Yup, that's the last aspect I was asking about, if chan 0 will always follow the same rule, so that there's not a confusing situation for newbies where they can hear someone but not see them

[17:10] Simon Linden: just avatars, attachments and ridden vehicles

[17:10] Michelle Resistance: Very nice.

[17:11] Ashiri Sands often sees driverless vehicles

[17:11] Simon Linden: Finally, one more change: avatars that are separated by more than 250m vertically aren't capable of seeing each other.

[17:11] Simon Linden: This also will give us a minor performance boost when regions are using skyboxes and platforms

[17:11] Michelle Resistance: Er.

[17:11] Michelle Resistance: Is that in all cases, or just with that setting enabled?

[17:11] Simon Linden: If things go smoothly, this would likely hit an RC channel in July and the full grid a few weeks later, but that's just a WAG at the schedule

[17:11] Simon Linden: That's all cases, Michelle

[17:11] Michelle Resistance: That's... not entirely desirable. x_x;

[17:12] NielArcher: so if my draw distance is set to 512, I can't see an avatar 256m above me?

[17:12] Jonathan Yap: If there is an uproar they will probably back out that change

[17:12] Kallista Destiny: well Pratically how many peopl have draw distance set Greate tht 256 anyway

[17:12] Michelle Resistance: I'm staff at an estate where people routinely run draw distances of 1024m or higher, and often are interacting with people separated by 300m vertically.

[17:12] Simon Linden: I'm definitely interested in hearing of the cases where that's really important

[17:13] Pauline Darkfury: Yeah, going by beta chat, we'd like that setting to be EM configurable via SimConsole, please, both in terms of max delta-Z and turning it on/off

[17:13] NielArcher: I have mine set to 256 normally. Occasionally to 512

[17:13] Kallista Destiny: I stay around 170 just for frame rate

[17:13] Simon Linden: if it was a larger number, say 512 or 1024, would it be simpler and more acceptable?

[17:13] Michelle Resistance: Yes.

[17:14] Michelle Resistance: Or, if it were overridable by the client.

[17:14] Kallista Destiny: Yeah 1K meters is a long wat to actually se a person with unaided eyes

[17:14] NielArcher: I'm having trouble understanding why it's not just down to your DD

[17:15] Simon Linden: it's much simpler to just use a single number, and easier to understand ... and possibly space out skyboxes

[17:15] Michelle Resistance: It seems like the use case you're thinking of, Simon, is really what would be served by having a separate draw distance slider for avatars.

[17:15] NielArcher: ahh I suppose it is, after the sim has sent relevant data

[17:15] Pauline Darkfury: Presumably sending the prim data to the client is already trimmed on the interest list if they are out of draw dist

[17:15] Pauline Darkfury: ?

[17:15] Andrew Linden: w

[17:15] Simon Linden: yes, there's trimming of data on both the server and viewer

[17:15] Pauline Darkfury: For the skybox case, Simon, it's tricky to have a single number

[17:16] Jane1 Bookmite: see you all soon,,,

[17:16] Pauline Darkfury: WIth a full region split into 4096s, 250m is about right for skybox separation

[17:16] Simon Linden: Take it easy, Jane

[17:16] Pauline Darkfury: If they are 1/4 sim, 1000 is about right

[17:17] NielArcher: 512 would work for me. I can't really think of any situation where 1024 would be useful. not given current size limitations on avatars

[17:17] Michelle Resistance: I'm very concerned about this idea, though.

[17:17] Michelle Resistance: It could also make administering a highly-vertically-occuped sim very hard.

[17:17] Simon Linden: could you send me a notecard and describe how you use things today, and where this would cause problems?

[17:18] Pauline Darkfury: Well, from beta chat, the admin need to see them on sim-wide radar isn't impacted

[17:18] Kallista Destiny: Thank you all for a most informative meeting

[17:18] Michelle Resistance: I suppose, but I feel like this isn't something that we can't speak for all the use cases of.

[17:18] Simon Linden: Thanks everyone for coming

[17:18] Pauline Darkfury: Thanks folks

[17:18] Pauline Darkfury: Have a good weekend

[17:18] NielArcher: likewise. thanks Andrew, Simon

[17:18] Jonathan Yap: Thank you Simon and Andrew

[17:19] Ashiri Sands: Thanky ou Simon & Andrew

[17:19] NielArcher: just about bed time for me ;-)

[17:19] Peewee Musytari: thanks andrew & simon ..good meeting :)

[17:19] Jonathan Yap: I'm about to find out if windlight clouds and glow are fixed in the shiny branch

[17:20] Andrew Linden: Thanks for coming everyone. Have a good weekend.