Simulator User Group/Transcripts/2013.07.23

From Second Life Wiki
Jump to: navigation, search


Prev 2013.07.16 Next 2013.07.30

List of Speakers

Baker Linden Davido Chrome DrFran Babcock
Gaius Tripsa Ima Mechanique Kennylex Luckless
Kitto Flora Lucia Nightfire Mona Eberhardt
Nalates Urriah Petr Topaz Qie Niangao
Rex Cronon Simon Linden Sit Test
Sovereign Engineer Talarus Luan Tiberious Neruda
Visitor Detector Whirly Fizzle Yuzuru Jewell
Zander Roxley


[12:01] Simon Linden: Hello everyone ... I can get started with the server update news

[12:01] Simon Linden: The forum post is here :

[12:02] Davido Chrome: Hav you automated the transcripts?

[12:02] Simon Linden: We updated the main channel this morning with the "Experience tools" project.

[12:02] Simon Linden: This really shouldn't affect anything, and unfortunately I don't have any updates or info on how this might be released as a feature or product

[12:03] Rex Cronon: owner use "llsl:Rex Cronon" for que use "chatLog" and go here>

[12:03] Visitor Detector: Returning visitor to your land: Yuzuru Jewell

[12:03] Davido Chrome: So it's out there, but not strictly usable?

[12:03] Simon Linden: The server-side appearance code has been doing well and the full grid rollout is likely soon

[12:04] Simon Linden: Yes, AlveKatt

[12:04] Visitor Detector: Returning visitor to your land: Sovereign Engineer

[12:05] Talarus Luan: O.o

[12:05] Simon Linden: Tomorrow we'll get the RCs updated. The maintenace branch on BlueSteel is the same, and the server-side appearance code will be on LeTigre and Magnum

[12:05] Talarus Luan: Why does everyone have to pick on the chicken?

[12:05] DrFran Babcock: another day, another pusher

[12:06] Sovereign Engineer: D:

[12:06] Simon Linden: I'm not monitoring the playground so send me an IM if anything gets out of hand :)

[12:06] Petr Topaz: mmm KFC

[12:07] Sovereign Engineer: \o/

[12:07] Simon Linden: That's it for release news ... Baker, did you have anything?

[12:07] DrFran Babcock: omg.

[12:07] Baker Linden: I'm still working on the viewer-side implementation of group bans

[12:08] Baker Linden: I'm currently deciding on the format of the data coming into the viewer, and adding it to the group manager subsystem in the viewer.

[12:08] Baker Linden: That's about it :)

[12:08] Visitor Detector: Returning visitor to your land: Whirly Fizzle

[12:08] Simon Linden: ok - the table's open for topics and questions then

[12:09] Talarus Luan: Any news on the interest list work Andrew was taking care of?

[12:10] Simon Linden: No - he was wrapping that up but it wasn't ready for release yet. I don't know if it's a possibility for next week or not

[12:10] Simon Linden: He was on vacation a bit so progress stopped during then

[12:12] Simon Linden: This might be a very quiet meeting :)

[12:12] Kitto Flora: No one's working on the serious problems... what's to ask?

[12:12] Kennylex Luckless: Btw; I has a convert material question. Is it possible to do an auto "undo" if you try to convert an object and it get to high LI? Now objects are returned and then it can be hard to restore them.

[12:13] Kennylex Luckless: Ie, if an object get to high LI it get back to original state instead of getting returned

[12:13] Davido Chrome: Can we expect there to get the world update problem for quick turns to come back when the interest list work rolls out?

[12:13] Davido Chrome: What Kenny said!

[12:14] Tiberious Neruda: yeah... I remember that was mentioned yesterday at the mesh meeting

[12:14] Simon Linden: Kennylex - so I understand it ... you add materials, this switches you to the new accounting which gives you a higher cost, it gets returned and you're kinda stuck?

[12:14] Nalates Urriah: Or a feature to estimate the LI when materials are added...

[12:15] Kennylex Luckless: I did miss that, but this is for all objects.

[12:15] Kennylex Luckless: I did try convert a build to convex hulls and it become around 35 000 LI.

[12:15] Rex Cronon: wow kennylex

[12:16] Tiberious Neruda: erf... that sucks

[12:16] Talarus Luan: That's... ridiculous.

[12:16] DrFran Babcock: where does one rez that ???

[12:16] Talarus Luan: In OpenSim. :P

[12:16] Tiberious Neruda: I don't think he -can-

[12:16] Rex Cronon: there r not enought prims in the sim for that:(

[12:16] Kennylex Luckless: To fix it you has to wear the build and eddit while wearing it

[12:16] Qie Niangao: right.

[12:16] Davido Chrome: Ah, that's clever.

[12:16] Lucia Nightfire: LL will have to fix before they allow parcel prim overflow, heh

[12:17] Qie Niangao: (Also, I swear that sometimes the LI *temporarily* goes super-critical, then settles down to a real number.)

[12:17] Simon Linden: hmm, yeah, that sounds pretty ugly

[12:17] Kennylex Luckless: There has to be some kind of warning.

[12:17] Tiberious Neruda: but this brings up an interesting question... what was the limit of 15,000 prims a sim originally set for?

[12:17] Davido Chrome: Closed Jira.

[12:17] Tiberious Neruda: like, why was that the limit?

[12:18] Rex Cronon: i think was for havok

[12:19] Simon Linden: I think it was a best guess at the time for a reasonable number ... given the variety of prims, it's not a real measure of how loaded or full the region is, but it needed some limit

[12:20] Davido Chrome: You still can make simcrashingly complex physics shapes with 15000 prims I gather.

[12:21] Kitto Flora: What is being changed for Materials? I look at the LL Wiki for Materials - I don't see any changes.

[12:21] Kennylex Luckless: But I think LI system is good, it has made me optimize mote, but it need to be safe and give a carrot to thise who do optimize

[12:21] Talarus Luan: The thing I keep coming back to is that the use of any of the new mesh/materials stuff is seriously discouraged by the unpredictability (and often outright lunacy) of the LI system.

[12:21] Lucia Nightfire: if you need physics to crash sims you're doing it wrong

[12:21] Tiberious Neruda: okay... then that begs the next logical question: now that we're WELL off of Havok 1, and into the 'LI era', why is 15,000 still the limit?

[12:21] Simon Linden: The new accounting is an attempt to get a more accurate measurement of "cost" but that turns out to be really complex ... and I know it's not perfect

[12:22] Tiberious Neruda: not perfect?

[12:22] Davido Chrome: Yeah. You really need safeguard from makiing an object unrezable by accident.

[12:22] Simon Linden: I know you don't think it's perfect

[12:22] Tiberious Neruda: not PERFECT?! No. It';s outright B-R-O-K-E-N

[12:22] Sovereign Engineer: <.,

[12:22] Kennylex Luckless: Is it not better that it is easy to understand rather than accurate?

[12:22] Sovereign Engineer has an object with an LI of 630,000

[12:22] Sovereign Engineer: o.o

[12:22] Whirly Fizzle: Wow lol

[12:23] Sovereign Engineer: or was it 640

[12:23] Lucia Nightfire: I think LL's next chance at acceptable "gamble" of changing the LI calc would be on top of increasing the 15000 LI limit

[12:23] Davido Chrome: With the new LI system you could lower the max to force people to use optimized builds, right? You can have soo much more detail.

[12:23] Zander Roxley: 630K

[12:23] Talarus Luan: Well, there's a bit of distance between "not perfect (but close)" and "not perfect (and mostly unusable)".

[12:23] Tiberious Neruda: the LI system, especially with mesh, actually PUNISHES you for making LOD models

[12:24] Lucia Nightfire: 12.5 million land impact linkset - <--beat this high score for a single object LI

[12:24] Sovereign Engineer: Is that a challenege?

[12:24] Sovereign Engineer: o.o

[12:24] Sovereign Engineer: :>

[12:24] Davido Chrome: Tiberious: Does that include when you make a prim physics shape for the Mesh model?

[12:24] Lucia Nightfire: idk, I wasn't even trying, just impressed

[12:24] Tiberious Neruda: yes, that includes a physics mesh

[12:25] Davido Chrome: Tiberious, you can lower the LI by making the physics shape from untortured prims rather than Mesh.

[12:25] Tiberious Neruda: in fact, the LI system encourages gaming the ever-loving HELL out of it by making one mesh for the High LOD, and using it for the top -3- slots

[12:26] Simon Linden: Uhm, I would like to point out that's not a "score" where you're trying to get bigger numbers ;)

[12:26] Tiberious Neruda: and generating the lowest 'cause nobody ever really sees that one

[12:26] Lucia Nightfire: nope, but fun nonetheless

[12:26] Simon Linden: but that said, it's pretty funny

[12:26] Sovereign Engineer is already thinking about how to get one above 12.5 mil

[12:26] Sovereign Engineer: o.o

[12:27] Talarus Luan: Funny for you, disheartening for me. :-/

[12:27] DrFran Babcock tries to make everything 1LI if it's going to be rezzed. Challenging

[12:27] Simon Linden: so was it the physics shape that pushed it up so high?

[12:27] Zander Roxley: 0.5

[12:27] Rex Cronon: imo all these weirdly huge LI are not really needed

[12:27] Whirly Fizzle: What happens if you wear that object Lucia? Do we all crash? :D

[12:27] Sovereign Engineer: uh

[12:27] Lucia Nightfire: no physics involved

[12:27] Sovereign Engineer: your framerate will be very sad.

[12:27] Sovereign Engineer: Especially toaster gpus :o

[12:27] Tiberious Neruda: can't rez it... it's over 15,000 LI

[12:27] Sovereign Engineer: you can wear it.

[12:27] Sovereign Engineer: silly.

[12:27] Lucia Nightfire: had I added physics it would probably have been astronomical

[12:27] Kennylex Luckless: If prims get to small it should automaticly become convex hull

[12:28] Sovereign Engineer: or rez it as temp

[12:28] Sovereign Engineer: o.o

[12:28] Whirly Fizzle: lol

[12:28] Davido Chrome: Or none.

[12:28] Kennylex Luckless: and even when holes become to small

[12:28] Lucia Nightfire: idk what my ARC would have been wearing it either, heh

[12:28] Rex Cronon: the LI number does not represend the actual nr of objects in a linked set. it represents how heave an linked set is

[12:29] Rex Cronon: heavy

[12:29] Rex Cronon: how heavy*

[12:29] Visitor Detector: Returning visitor to your land: JayR Cela

[12:29] Talarus Luan: The LI number is an amorphous, opaque calculation

[12:29] Tiberious Neruda: but, this just goes to show why the LI system is so bad as it stands

[12:29] Kennylex Luckless: I wear 15 000 LI object now but ARC not go up so much, but it is for things you wear has no phys shape.

[12:29] Lucia Nightfire: one thing I frown on with LI is, I wish it could be displayed along with prim counts in the prim owner's list

[12:29] Davido Chrome: Though if you can fool the LI system by only simplifying the last LOD that must be considered a bug, right?

[12:29] Talarus Luan: With prim builds, it is simple and transparent.

[12:29] Simon Linden: At a really basic level, the system is designed to give feedback on doing things that are really costly for the region ... so it definitely doesn't like something you're doing to it ... too many triangles, concave surfaces, that kind of stuff

[12:30] Tiberious Neruda: I mean, don't get me wrong...

[12:30] Rex Cronon: if i set a linked set to pahtom its land impact should be zero

[12:30] Tiberious Neruda: the idea is pretty damned good... it's just the implementation screwed the pooch

[12:30] Lucia Nightfire: Davido, I wouldn't call it a bug, I'd call it breathing room

[12:31] Talarus Luan: Well, we've already seen that is irregular.

[12:31] Davido Chrome: How so?

[12:31] Lucia Nightfire: as many content creators rely on that explicitly I would think

[12:32] Davido Chrome: Rex, you have to consider that LI are also part of SL economy, I think.

[12:32] Rex Cronon: that its is a bizzare and unrealistic economy:(

[12:32] Lucia Nightfire: all of my mesh creations that don't rely on physics structure I upload with either the last LOD or last two LOD values at 1 or 0 to get the lowest LI while maintaining decent LOD by distance and size

[12:32] Tiberious Neruda: it allows for being gamed to hell and back, and puts seriously unreasonable expectations on mesh creators

[12:33] Talarus Luan: Phantom just zeroes out the physics cost.. there's still simulation costs for rendering and transfer.

[12:33] Tiberious Neruda: I mean, if I make an object that's only 62 triangles at highest LOD, why do I need to lower THAT?

[12:33] Rex Cronon: escuse me. the server doesn't render the prims

[12:33] Visitor Detector: First time visitor to your land: Zidonuke Ghost

[12:33] Lucia Nightfire: it sort of makes sense to me anyway LOD levels acting as multipliers or exponentials

[12:34] Ima Mechanique: no, but it does send the geometry

[12:34] Davido Chrome: Simon, the LI system is supposed to reward if you have made incremental simplifications in the LODs, right?

[12:34] Talarus Luan: No, it doesn't, but 1) it does transfer them, and 2) render cost figures into LI.

[12:34] Rex Cronon: actually the data is located elsewhere

[12:34] Simon Linden: I don't know the details on the LOD effect

[12:34] Talarus Luan: It still knows about them, because it has to run them through the interest list.

[12:35] Kennylex Luckless: I do not like that a mesh as a cube that is optimized shape has hither LI in large scale than a sculptie

[12:35] Lucia Nightfire: certain stuff needs to maintain it LOD throughout all distances though, like high rez frequency mesh

[12:35] Simon Linden: THere's some basic info here :

[12:35] Rex Cronon: and is not like a linked set has a few gibabytes of data

[12:35] Simon Linden: and it shows how you can see the download, physics and display factors

[12:35] Lucia Nightfire: my shields have to have full LOD or whenever a plate rezzes its busy loading LODs

[12:35] Rex Cronon: if it does than it is a monster

[12:35] Davido Chrome: Lucy, I think you need to figure out other ways to work around such limitations.

[12:36] Lucia Nightfire: there are none in that regard

[12:36] Tiberious Neruda: that's only 62 triangles

[12:36] Tiberious Neruda: why does that -need- lowered any further?

[12:36] Lucia Nightfire: but as I was saying about the last two LOD levels beign low it allows for complex mesh without high LI cost

[12:37] Rex Cronon: to rez very, very, very, very fast

[12:37] Rex Cronon: faster than a bullet that comes towards u at full speed

[12:37] Lucia Nightfire: I have stuff I uploaded thats 30 - 40k tri's taht uses only a few LI, where as if I maintained that LOD throughout the levels it would have been several thousand LI

[12:37] Tiberious Neruda: oh that, I have no doubt

[12:37] Tiberious Neruda: but 30,000 triangles is, frankly, overkill

[12:38] Lucia Nightfire: true, but my app required it explicitly

[12:38] Lucia Nightfire: and I don't use many that use that level, heh

[12:38] Davido Chrome: Rex, the physics shape doesn't have LODs, so the shield would need to be fully rezzed to work?

[12:38] Lucia Nightfire: I know how to optimize in normal apps

[12:39] Davido Chrome: Wouldnot*

[12:39] Rex Cronon: depends what objects u use

[12:39] Tiberious Neruda: but, when an object is only up to about 640 triangles at highest LOD, and you provide LOD meshes for the other 3 slots that do lower... why is that punished?

[12:39] Kennylex Luckless: This bear is free to copy, try to optimize it by give some patrs "none" as physical shape.

[12:40] Davido Chrome: No Mesh in that?

[12:40] Talarus Luan: Oh, quick question, Simon.. would it be possible to find out the mathematical difference between llSitTarget positioning of avatars and llSetLinkPrimitiveParams positioning of same? Strife and I have been trying to make a more accurate UpdateSitTarget.

[12:40] Kennylex Luckless: No mesh, mesh are for losers.

[12:40] Lucia Nightfire: I would rez a 5ht iteration Menger Sponge here, but its 600k tri's and may lag everyone, lol

[12:40] Lucia Nightfire: *5th

[12:40] Kennylex Luckless: Do not look at mesh I did and have here :-)

[12:41] LagMeterVer4e: Prims Are Getting Low At Location: Linden Village - Simon Linden

[12:41] Davido Chrome: Lucy, Oooh, Is it a fractal?!

[12:41] Talarus Luan: Mmmmmmmenger Sponges :D

[12:41] Talarus Luan: Yup

[12:41] Lucia Nightfire: idt its a fractal but its along the lines

[12:41] Talarus Luan: It is

[12:41] Davido Chrome: Lucy, you can't tell and not rez!

[12:41] Simon Linden: Talarus ... you mean the difference between where a sit position gets set and a prim position?

[12:41] Gaius Tripsa: We're not spongeworthy?

[12:41] Talarus Luan: Simon: yes.

[12:42] Simon Linden: I'm not sure what you're asking but it sounds more interesting than the LI calculations :P

[12:42] Talarus Luan: As it relates to avatar positioning.

[12:42] Tiberious Neruda: and this is the other object...

[12:42] Tiberious Neruda: it's only like 640 triangles at highest LOD

[12:42] Talarus Luan: I've asked about it in the past, but never have been able to get someone to look at the code and help answer it.

[12:42] Tiberious Neruda: and if you look, it has decent LODs

[12:42] DrFran Babcock: and 2 LI

[12:43] Tiberious Neruda: exactly

[12:43] Simon Linden: I'm guessing that would depend on the prim size and shape ... you end up sitting on a surface, which is some distance from the location

[12:43] Talarus Luan: We have an experimental function which approximates it.

[12:43] Qie Niangao: Talarus, what makes you think that the current approximations aren't precise?

[12:43] Tiberious Neruda: the shiny one was 'cheated'

[12:43] Simon Linden: ... prim location, or "center" even if that's not really in the center

[12:43] Davido Chrome: Simon, I don't think that's true for scripted sits?

[12:43] Tiberious Neruda: and it's 1.7 download weight

[12:43] Talarus Luan: Empirical testing, Qie

[12:43] Simon Linden: by scripted you mean with a sit position that's been set, right?

[12:44] Tiberious Neruda: so, having the custom LODs was actually punished

[12:44] Davido Chrome: Isn't it allways relative to the prims center?

[12:44] Talarus Luan: Depending on orientation and avatar height, it can be off by up to several centimeters

[12:44] Qie Niangao: I wonder, because it's easy to test... but then it was by just such testing that the current approximation was derived.

[12:44] Talarus Luan: I've discussed it with you in the past. :)

[12:45] Qie Niangao: heh, that was a *long* time ago.

[12:45] Talarus Luan: Yup, Simon

[12:45] Tiberious Neruda: 'cheated' (high for top 3 LODs, and a generated Lowest): 1.7

Custom LODs: 2.3

[12:45] Talarus Luan: Yus. This is SL. Issues have long lives. :P

[12:45] Tiberious Neruda: that's why I say the LI system punishes you for trying

[12:45] Davido Chrome: Also, A fat butt will interesect if you were in a thin avatar when testing the sit position. The center of the avatar is the hip somewhere, I think?

[12:46] Lucia Nightfire: I desperately tried to read taht slowly, lol

[12:46] Kennylex Luckless: It is hare to use LI, but it can be rewarding.

[12:46] Talarus Luan: Basically, there is a mathematical discrepancy between the sit target and where the avatar is positioned via llSLPP

[12:46] Tiberious Neruda: what I'm saying is the LI system should reward you for making the LODs, even if they're not the drastic cuts LL expects

[12:47] Kennylex Luckless: and materials are awesome, I has some in the Borg ship. It shall be free to copy too

[12:47] Simon Linden: yeah, I can imagine AV size might come into play for that ... I don't know what exact part gets tied to the position, but assuming it's in the AV center / core you wouldn't match the outside of the shape

[12:47] Talarus Luan: The UpdateSitTarget function in the wiki is an approximation, and we've refined it a bit, but it is still not 100% accurate.

[12:47] Sit Test: Hello, Avatar!

[12:47] Qie Niangao: Tal, I vaguely recall somebody claiming that the leg length of the avatar shape, specifically, would affect how the sit offset worked. This back when Lex Neva and company were first trying to derive the equation we all use now.

[12:48] Talarus Luan: Escort DeFarge, I think

[12:48] Qie Niangao: ah, yeah, I think that's correct.

[12:48] Rex Cronon: using LI is like being slowly tortured. i guess that is good if u like bdsm:)

[12:49] Davido Chrome: I can't find the wiki page for update sit target...

[12:49] Kitto Flora:

[12:49] Talarus Luan: Anyway, point is, there is a function of some kind which will give a 100% translation, but it really needs some LL input from someone looking at the code.

[12:49] Talarus Luan: Yeah, it's in llSitTarget

[12:51] Davido Chrome: Oh, so it's a safeguard function of sorts?

[12:51] Simon Linden: I suspect the server side is actually pretty simple, but what you want is the right viewer result where people look seated properly

[12:52] Talarus Luan: Well, no, it's all server-side

[12:52] Qie Niangao: yeah, it's not simply a question of appearance...

[12:52] Qie Niangao: it's kind of... repeatability.

[12:53] Talarus Luan: I don't care where they are seated, I just want to make sure I put the same coords into llSLPP that match the ones in llSitTarget

[12:53] Talarus Luan: As it stands, there is approximately a 0.4m discrepancy in the Z axis.

[12:54] Lucia Nightfire: iirc does that get off if they have shoes on or not?

[12:54] Lucia Nightfire: shows you how much I work with visuals and sittarget, heh

[12:54] Talarus Luan: and that approximation varies based on at least one factor.

[12:54] Talarus Luan: Supposedly it does, but I haven't tested that part.

[12:55] Simon Linden: What are you setting via llSLPP?

[12:55] Simon Linden: prim pos?

[12:55] Talarus Luan: The avatar position

[12:55] Talarus Luan: as a link on the link set

[12:55] Simon Linden: ok, tell me the PRIM_WHATEVER flag you're using

[12:55] Talarus Luan: PRIM_POSITION

[12:55] Davido Chrome: In what context does this thing appear?

[12:56] Lucia Nightfire: lmao, we need PRIM_WHATEVER now

[12:56] Talarus Luan: Furniture, mostly

[12:57] Davido Chrome: I know you can offset the avatar in teh making of the pose. Could that be it?

[12:57] Talarus Luan: If someone sits on a sit target, and is moved with PRIM_POSITION to the same offset, they will be off by some amount of error.

[12:57] Simon Linden: ok ... so that sets the offset for the center point of a given prim. That won't be the same as a sit position, right?

[12:57] Davido Chrome: Ooh.

[12:57] Davido Chrome: Could it be it gets Prim position plus Sit target in the new prim?

[12:58] Simon Linden: ah, I think I get it ... so you've set the set target. You then move the prim, and the sit changes? Is that the problem?

[12:58] Talarus Luan: Not moving any prims, moving the avatar with PRIM_POSITION

[12:58] Talarus Luan: the seated avatar

[12:59] Lucia Nightfire: my global pos versus teh sittarget here is off by .35

[12:59] Kennylex Luckless: Oh, a luxyry moded.

[12:59] Talarus Luan: When an avatar sits on a linkset, they become a link in the linkset that you can move around with PRIM_POSITION

[12:59] Kitto Flora: ... and nobody yet mentioned the pose construction ;)

[13:00] Talarus Luan: The position vector you use is off from the sit target by some factor, based on the at least the avatar height.

[13:00] Rex Cronon: don't animations also change ave position?

[13:00] Simon Linden: they can, Rex

[13:00] Talarus Luan: They can, but that is an additional consideration

[13:00] Rex Cronon: maybe is the animation that introduces the offset

[13:01] Rex Cronon: ?

[13:01] Talarus Luan: I just would like the base translation function to be accurate. I can then deal with animation offsets.

[13:01] Davido Chrome: So basically you can reset it?

[13:01] Talarus Luan: No, because I am using the same animation in both cases to test it

[13:01] Davido Chrome: Without unsitting?

[13:01] Davido Chrome: Can't*

[13:01] Kennylex Luckless: See later, I has to go.

[13:01] Rex Cronon: tc all those leaving

[13:02] Visitor Detector: Returning visitor to your land: Nalates Urriah

[13:02] Simon Linden: You get different results with different sized AVs, right?

[13:02] Talarus Luan: Correct.

[13:02] Talarus Luan: That's why there is a llGetAgentSize in UpdateSitTarget

[13:03] Rex Cronon: btw. all those of u that were late. u can see the transcript now. for owner use "llsl:Rex Cronon" for que use "chatLog" and go here>

[13:03] Talarus Luan: But there are also some other factors involved

[13:03] Davido Chrome: Talarus, what do you use this function for? Customizable sit position?

[13:03] Talarus Luan: and I would like to know what they are, and if I can even include them in a translation function

[13:03] Yuzuru Jewell: Rex, I want to see it. but I can't.

[13:04] Talarus Luan: Yes. it is a poseball-less furniture system

[13:04] Davido Chrome: Rex, did you log mondays meeting?

[13:04] Rex Cronon: i launched my site today. nothing happend monday

[13:04] Davido Chrome: Nothing?!

[13:04] Baker Linden: take care everyone! Have a good week! See some of you Thursday!

[13:05] Rex Cronon: tc baker

[13:05] Yuzuru Jewell: Thank you, Backer.

[13:05] Rex Cronon: yuzuru. what u mean u can't?

[13:05] Talarus Luan: Anyway, if you can look into it and help out any, we would appreciate it, Simon. :)

[13:05] Yuzuru Jewell: I can't see the chat log of this meeting.

[13:06] Yuzuru Jewell: I am late for 10min.

[13:06] Lucia Nightfire: yeah, SLPP is accurate, sittarget sux, heh

[13:06] Simon Linden: I see one factor in the code - it's shifting 0.75 * the AVs x direction scale

[13:06] Rex Cronon: yuzuru. did u go to

[13:07] Yuzuru Jewell: yes

[13:07] Rex Cronon: exclude the question mark

[13:07] Talarus Luan: x?

[13:07] Yuzuru Jewell: yes

[13:07] Yuzuru Jewell: there is no log.

[13:08] Rex Cronon: for owner textbox use(llsl:Rex Cronon), for que textbox use(chatLog)

[13:08] Talarus Luan: tried it, Rex.. get "Que is missing"

[13:08] Simon Linden: yes, the x component from llGetAgentSize()

[13:08] Talarus Luan: huh.. interesting

[13:08] Davido Chrome: Could be that Sittarget and PRIM_POSITION uses different origin values? One might use the pose bone position value while the other uses collision box center, perhaps?

[13:08] Yuzuru Jewell: Oh I can see it!

[13:08] Yuzuru Jewell: Thank you!

[13:08] Talarus Luan: Well, if it was just a simple offset, that would be easy to calculate.

[13:08] Rex Cronon: :)

[13:09] Talarus Luan: But it changes based on several factors, only one of them known.

[13:09] Simon Linden: That's the one calculation I can see in the code during a sit that looks like it's using the AV size

[13:10] Talarus Luan: and is x, rather than z? Strange

[13:10] Yuzuru Jewell: Rex, It is very useful!

[13:10] Lucia Nightfire: it uses it so your size not hovering over the prim like you were in a public bathroom, lol

[13:10] Simon Linden: oh, never mind, I'm looking in the wrong place .. that's un-sitting code

[13:10] Simon Linden: sorry

[13:10] Rex Cronon: u r welcome:)

[13:12] Talarus Luan: So far, I have sampled avs from ~1.16m to ~2.45m tall, and created a second-order curve function which approximates the variations better than the current UpdateSitTarget, but it is still an approximation.

[13:13] Rex Cronon: simon. i u guys had a uml diagram would be relativley easy to find things without getting lost in code:)

[13:13] Talarus Luan: It is exact for the three control points I use to generate the curve, but still off for all others to one degree or another.

[13:14] Simon Linden: I'm not seeing any other offsets using the AV size, but it must be buried somewhere

[13:14] Davido Chrome: Could possibly be that the avatars intersecting physics shape pushes it out a bit?

[13:14] Simon Linden: I'm sorry but I have to get going

[13:14] Davido Chrome: Thank you Simon

[13:14] Mona Eberhardt: bye Simon

[13:14] Davido Chrome: Thatnk you Simon!

[13:14] Rex Cronon: it could be a constant

[13:15] Yuzuru Jewell: Thank you, Simon.

[13:15] Simon Linden: Thanks for comig today


Prev 2013.07.16 Next 2013.07.30