User:Andrew Linden/Office Hours/2008 04 08

From Second Life Wiki
Jump to: navigation, search

Transcript of Andrew Linden's office hours:

[11:01] Arawn Spitteler: Afterlife, requires Age Verification.
[11:02] Bato Brendel: sup andrew
[11:02] Les White: hiya Andrew
[11:02] Kitto Flora: Oh
[11:02] Andrew Linden: hey everyone
[11:02] Kitto Flora: That sim got set that way too?
[11:02] Scripting Student: hiya
[11:02] Rex Cronon: hi everybody
[11:02] Arawn Spitteler is old enough, for the Adult Grid, but is much too young to die.
[11:02] Kitto Flora: I found that WIsh was suddenly set Age Verify a couple days back
[11:02] Strife Onizuka: do we have an agenda?
[11:02] Kitto Flora: Hi Andrew
[11:03] Bato Brendel: Im personally just going down the list of 'physics' 1.20 server marked and checking them out
[11:03] Andrew Linden: No real agenda. I'll give a summary of the Havok4 bug fixes, and any announcements I can think of, then we'll open the table for whatever.
[11:03] Andrew Linden: We're currently testing some bug fixes that we'd like to deploy today
[11:04] Strife Onizuka nods
[11:04] Bato Brendel: sim FPS?
[11:04] Bato Brendel: for the memory allocation?
[11:04] Andrew Linden: I'm trying to get one particular bug into this release
[11:04] Les White: *fix
[11:04] Les White: hehe
[11:04] Andrew Linden: right... bug *fix*
[11:05] Andrew Linden: so I was contemplating bailing on these office hours, but decided against it ;-)
[11:05] Bato Brendel: well atleast you dont have an angry mob with torches all standing around 8)
[11:05] Andrew Linden: First of all, let me tell you what is NOT fixed...
[11:06] Andrew Linden: hovering vehicles still broken, jerkey camera motion are not fixed
[11:06] Andrew Linden: llTargetOmega() getting mysteriously reset or lost (for visuals) not fixed
[11:07] Andrew Linden: Ok, so what is fixed...
[11:07] Andrew Linden: lemme consult some notes...
[11:07] Les White: what's broken with hover vehicles?
[11:07] Bato Brendel: lemme cap that one
[11:08] Gaius Goodliffe: They don't hover at the height you specify for one...
[11:08] Bato Brendel: is a good example of one of the problems as well as hoverheight
[11:08] Andrew Linden: Er... buoyant hover vehicles fall from the sky
[11:08] Andrew Linden: So fixed... SVC-1786 holding spacebar down for slow walk
[11:09] Kitto Flora: Hover problem is getting to be a pain. Lots of complains, all my 'boat' products are 'swamped'.
[11:09] Andrew Linden: a scripted sim crasher, and some avatar-sit bug
[11:10] Gaius Goodliffe: Kitto: your boats are hoving *lower* than specified?
[11:10] Andrew Linden: That's it really, there were a few others that we're trying to get in.
[11:10] Kitto Flora: Yes - all at 0.5M low
[11:10] Gaius Goodliffe: Bizarre, mine are hovering precisely 0.218m higher than specified.
[11:10] Andrew Linden: boats too low is not fixed
[11:10] Bato Brendel: yes as well a few of my boats seem to be actually 'high'
[11:10] Andrew Linden: yeah, it is a mystery as to why some boats are high and others are low
[11:11] Kitto Flora: OK Andrew, whats the plan for fixing Hover? It was great a couple weeks back, before main grid deploy.
[11:11] Rex Cronon: could it be that different types of prims are used?
[11:11] Andrew Linden: I need to dig into that, but have some other bug that is currently kicking my @$$
[11:11] Andrew Linden: No, prim types should not influence that behavior.
[11:11] Bato Brendel wonders as well if the 'water height' could be effecting this as well "boat ref."
[11:12] Andrew Linden: Ok... do you have any announcements Simon or Sidewinder?
[11:12] Andrew Linden: Otherwise, we'll open the table.
[11:12] Rex Cronon: how about where the position of firts linked object in the set?
[11:12] Sidewinder Linden: andrew - just checking in - is that internal issue you've been wrestling with close - think we'll pass qa on it today?
[11:12] Andrew Linden: no Sidewinder, it is not close.
[11:13] Sidewinder Linden: oof
[11:13] Sidewinder Linden: ok well we have a build that we hope to get out "soon", but it depends on a couple of very difficult resolutions
[11:13] Sidewinder Linden: it should take care of a few issues that people really care about from the big deploy, including the open space memory issue
[11:13] Simon Linden: Since you mentioned hovering, fwiw I fixed some problems with llSetHoverHeight and llGroundRepel and how they are dealing with water. Isn't in the next update, however as it just missed the deadline
[11:13] Sidewinder Linden: sry if andrew already mentioned this - been multi-threading
[11:14] Sidewinder Linden: ahh cool simon
[11:14] Bato Brendel: yes, well sidewinder was present yesterday the memory allocation for voids/open sims seems to be working in the beta. I did a pretty hard stress test on one of the voids on the beta
[11:14] Creem Pye: Gaius, are your ships hovering relative to absolute height?
[11:14] Kitto Flora: Great Simon! But how much delayed does that mean?
[11:14] Les White: (would love a global hover option for normal dynamic objects)
[11:14] Sidewinder Linden: great bato
[11:14] Sidewinder Linden: kitto - we'll have to see - hopefully only a week after this build goes out
[11:14] Gaius Goodliffe: Nope. Relative to the water height.
[11:14] Creem Pye: ah
[11:14] Sidewinder Linden: we are trying to do a release per week if we can pass qa for them
[11:14] Sidewinder Linden: for the next couple of weeks
[11:15] Kitto Flora: Any chance of putting on the BLOG that boats are borked, and please wait a week?
[11:15] Sidewinder Linden: there are a couple of issues that i've been hearing a lot about - we do know about them and will dive in on them soon
[11:15] Andrew Linden: Oh right, void region memory footprint was fixed... That was part of a bug that Simon worked on that had another part that is not yet fixed.
[11:15] Sidewinder Linden: i'll put that up on this release
[11:15] Sidewinder Linden: kitto
[11:15] Sidewinder Linden: btw NOT all boats are borked ;)
[11:15] Sidewinder Linden: just like not all cars are, regardless of what some folks wanna say ;)
[11:15] Gaius Goodliffe nods.
[11:15] Sidewinder Linden: but yes there are issues
[11:15] Creem Pye wonders if setting the hover height to absolute, with an offset of llWater() would help
[11:15] Simon Linden: And this fix was not in the VEHICLE parameters, just those LSL functions
[11:15] Arawn Spitteler: Are there any working trains?
[11:15] Sidewinder Linden: there are a few things that i've been hearing a lot about -
[11:16] Nber Medici: room for another at the table?
[11:16] Sidewinder Linden: 1) llTargetOmega objects stopping spinning in various circumstances
[11:16] Andrew Linden: Nber, grab the empty seat
[11:16] Simon Linden: Nber - sure, click and sit on the empty chair
[11:16] Sidewinder Linden: we have many cases of this to work with
[11:16] Sidewinder Linden: it's just a matter of getting past the current issues to start heavy on thatone
[11:17] Kitto Flora: All my steam trains, monorail, swan, boat, railcar are working, with the exception of HOVER HEIGHT, and a problem with collision distance
[11:17] Sidewinder Linden: 2) performance issues - most of these turn out to be individual scripts that do strange things on h4 vs h1, like some fish rezzers and those interpenetrating tori wave generators
[11:17] Sidewinder Linden: ok
[11:17] Sidewinder Linden: i heard a report that the linden trains were broken... but haven't had the chance to check it out yet
[11:17] Sidewinder Linden: 3) bumping over seams....
[11:17] Creem Pye: speaking of performance issues, I tried following my own suggestion and created an invisible "bounding box" around a car of mine. That managed to reduce the physics time to ~60% of what it was before, but the physics time was still maybe 4x higher than only having the bounding box as a 1-prim vehicle
[11:18] Sidewinder Linden: there's some real angst about this one, in various courts, but it's not an easy problem
[11:18] Nber Medici: bumping over seams!!!
[11:18] Kitto Flora: Linden Trains are mechanical... they may have issues
[11:18] Nber Medici: yeah :)
[11:18] Bato Brendel: I havent heard any reports of the GSLR train being borked, I can check on that later.
[11:18] Sidewinder Linden: some vehicle folks, and a number of sports folks are having big issues with that
[11:18] Arawn Spitteler wants to ask Jack Linden how important Rails are, since they were vital to Disneyland.
[11:18] Sidewinder Linden: these are just three that we already know about in some detail - wanted you to know that we know they're there, and are trying to get to them after the current fixes
[11:19] Nber Medici: May I ask a Q?
[11:19] Kitto Flora: GSLR train was working last I looked - but it may have been worked on in Beta.
[11:19] Sidewinder Linden: sure nber
[11:19] Sidewinder Linden: btw i'm standing only 'cause i hvae to jump in a cuople of mins - sorry
[11:19] Nber Medici: I am here on behalf of golf and the bumping over seams issues
[11:19] Sidewinder Linden: yes...
[11:19] Nber Medici: Bato and ZAT have posted those to JIRA
[11:19] Nber Medici: may I ask status?
[11:19] Andrew Linden: bumping over seams is known, and not yet fixed. Waht is the golf issue?
[11:19] Bato Brendel: it worked in the beta since perwinkle was there but it runs on a HTML location reporter for when it hangs up so I never ran them in there. but I can check it out
[11:20] Nber Medici: It has broken golf totally.
[11:20] Nber Medici: can't putt
[11:20] Andrew Linden: More details please.
[11:20] Sidewinder Linden: it's another seam bumpging problem - golf balls bounce going over prim seams
[11:20] Sidewinder Linden: *bumping
[11:20] Andrew Linden: oh
[11:20] Bato Brendel: yeah its a seam issue Andrew
[11:20] Andrew Linden: bummer
[11:21] Nber Medici: IT is a real financial problem for MarkTwain White and me.
[11:21] Arawn Spitteler is wondering about a fly-away at borders issue, has that been noticed?
[11:21] Nber Medici: It is
[11:21] Bato Brendel: even when the BBox's are flawlessly edged (irrelative) to prim posititon.
[11:21] Andrew Linden: At the moment there is special magic for vehicles that tries to compensate and fix the bumping on seams problem
[11:21] Andrew Linden: however it is not on by default for all objects
[11:21] Strife Onizuka: status flag possability?
[11:22] Andrew Linden: the cost of putting it on all objects would be prohibitive... best to move to a new physics engine that doesn't suffer from the problem
[11:22] Andrew Linden: and the special magic for vehicles needs some tweaking -- some vehicles still hit the prim seams hard
[11:22] Sidewinder Linden: hmm maybe we can talk about this internally... i would think we could come up with something to allow this to work for specific classes of objects somehow
[11:22] Nber Medici: Here is the Jira for golf
[11:23] Sidewinder Linden: as strife mentioned... maybe a parm that says "make this prim seam safe" or something to that effect?
[11:23] Nber Medici: well here is the big issue
[11:23] Bato Brendel: not possible for the whole golf thing the coder is MIA
[11:23] Nber Medici: and I will be the first to say
[11:23] Nber Medici: LOL
[11:23] Arawn Spitteler: Would it work, simply to declare them sleds?
[11:23] Sidewinder Linden: ruh roh
[11:23] Nber Medici: yest what Bato said
[11:23] Nber Medici: the scripter for golf has not been inworld for over 6 months
[11:24] Nber Medici: and does not respond to queries
[11:24] Andrew Linden: when I was working with the Havok engineers they mentioned something about a "seam welder" utility which could be used for fixing the prim seam problem
[11:25] Strife Onizuka: I have an obscure question, the fudge factor used to increase the probability of detecting collisions and causes objects to appear not to touch, for what percentage of collisions is the fudge factor neccessary for detection? (ballpark figure will do)
[11:25] Andrew Linden: however from what little I gathered about it, it was usually used for very specific content objects in most Havok-empowered games
[11:25] Andrew Linden: we might be able to bend it to our will for most moving objects, but I don't know what the costs or tradeoffs would be
[11:26] Andrew Linden: Strife, the fudge factor you're talking about is the "collision tolerance" ?
[11:26] Strife Onizuka: yes
[11:26] Andrew Linden: the collision tolerance is a fixed size 10 cm
[11:26] Kitto Flora: Is it know why golf balls (and bikes?) suffer from seam bumps, but my vehicles seem not to?
[11:26] Bato Brendel: well andrew would it be correct is saying that the bounding boxes for the physics seems to be 'pin cushioned' ie. the edges are stretched out beyond the actually prim size?
[11:27] Les White: depends on shape and speed, kitto
[11:27] Andrew Linden: Bato, I think you're seeing the "collision tolerance" effect of the Havok physics engine.
[11:27] Bato Brendel: yes, the larger the mass and size the less noticable is it.
[11:27] Bato Brendel nods
[11:27] Kitto Flora: OK
[11:27] Strife Onizuka: i know the size is fixed and i know it's there to increase the probability of detecting collisions for fast moving small objects and for detecting glancing blows
[11:27] Andrew Linden: The actual shapes are correct, but Havok says two objects are actually colliding when within that tolerance
[11:28] Gaius Goodliffe: Kitty: " Andrew Linden: At the moment there is special magic for vehicles that tries to compensate and fix the bumping on seams problem"
[11:28] Strife Onizuka: i'm just curious about the percentage of collisions that depend on it to be registered
[11:28] Nber Medici: The golf balls certainly react dramatically
[11:28] Gaius Goodliffe: /sorry, Kitto
[11:28] Andrew Linden: the reason they do that is to reduce the "penetration events at almost all costs"
[11:28] Sidewinder Linden: i am going to have to go.. sorry to have to jump - tues is rough schedule-wise... see you guys and thanks all :)
[11:28] Les White: see you sidewinder
[11:28] Bato Brendel: ie. keep them from getting sucked in between prims
[11:28] Rex Cronon: bye sidewinder
[11:29] Andrew Linden: Strife, I don't understand your questions, but I think it is because you're making assumptions about how the physics engine actually works. There are no "percentages" that I know about.
[11:29] Kitto Flora: I am getting complaints about collision events that did not happen in H1. Its like the H4 collision distance is a litle larger. Is that the case?
[11:29] Bato Brendel: very much
[11:29] Les White: looks like it is to me.
[11:30] Andrew Linden: Kitto, there are some other objects that appear to have that problem. The collision tolerance has not changed, but some of the shapes have...
[11:30] Bato Brendel: its very noticable when a hovering avatar gets even close to a moving 'water vechicle'
[11:30] Nber Medici: Well I know there are many things on the H4 agenda... but golf is important to us.
[11:30] Strife Onizuka: your previous statement about avoiding penetration makes me rethink my question
[11:30] Creem Pye: I've heard similar reports with "movelocks", which are attachments with a strong llMoveToTarget() to prevent an avatar from moving
[11:30] Andrew Linden: Symmetric spheres and cylinders are implicit, and may have different triangles.
[11:30] Kitto Flora: Andrew - In my case - it is a funny shape prim.
[11:30] Creem Pye: they're causing lots of collisions in H4
[11:31] Rex Cronon: right creem
[11:31] Strife Onizuka: movelocks are not new, they were first pioneered as a defense against orbitors and push weapons
[11:31] Zinfo Version 4.0
[11:31] Bato Brendel: its not movelock related strife with a hovering avy
[11:31] Creem Pye: maybe we don't really need movelocks anymore =)
[11:31] Rex Cronon: u do
[11:31] Bato Brendel: tried it with it on and off same effect,
[11:32] Rex Cronon: try having someone drive over u with a car
[11:32] Bato Brendel: yeah you still get pushed
[11:33] Bato Brendel: but like in ref to boats, you can be moving 1m/s and normally you would 'slowly' push the avy out of the way not anymore, you basically stop dead
[11:33] Creem Pye: mm but somebody was saying last week that their yachts couldn't push an avatar out of the way?
[11:33] Bato Brendel: yep
[11:33] Bato Brendel: you cant
[11:33] Bato Brendel: you stop the boat
[11:33] Kitto Flora: I'm seeing collision even raised when the range is > 0.13M
[11:33] Nber Medici: exactly.. another problem
[11:33] Andrew Linden: This is hovering avatars that collide with boats?
[11:33] Andrew Linden: Stop the boats?
[11:34] Creem Pye: maybe it's related to linear_motor_timescale (whether you can push an avatar or not)
[11:34] Bato Brendel: just gotta be close yes andrew I got a jira for it
[11:34] Bato Brendel:
[11:34] Andrew Linden: That problem is not related to the boat so much as the avatar motion code.
[11:34] Andrew Linden: It is probably fixable.
[11:34] Bato Brendel nods
[11:35] Les White: i dont see that with hover vehicles myself. the avs go flying
[11:35] Nber Medici: not with boats
[11:35] Nber Medici: they stop dead
[11:35] Bato Brendel: normal prims as well
[11:35] Bato Brendel: no vechicle code
[11:36] Andrew Linden: Bato, you mean "when non-vehicle prims hit avtars the prims stop moving" ?
[11:36] Bato Brendel: dont even gotta hit it just be 'close'
[11:36] Nber Medici: Bato is correct.
[11:37] Les White: i noticed when you set vehicle type (false) you get a dynamic object without friction. slides forever
[11:37] Andrew Linden: Les, I fixed that zero friction bug, but I don't think it made it into this next release.
[11:37] Gaius Goodliffe: FALSE == VEHICLE_TYPE_NONE ?
[11:37] Les White: yay :)
[11:37] Les White: though it could be handy
[11:37] Nber Medici: I can't tell what the status of the Linden response is to the 2006 JIRA. Can I get a read on that?
[11:38] Les White: false should remove vehicole params and make a normal dynamic object again
[11:38] Andrew Linden: Yeah, I would expect a few people to try to take advantage of that misfeature. Their content will break when the fix goes out.l
[11:38] Bato Brendel: was gonna say I was testing friction co-efficients for different MAT_TYPES for just plane jane physical prims an I could see the differences in it
[11:38] Gaius Goodliffe: That's what setting VEHICLE_TYPE_NONE should do. I assume it has the same value but I never checked.
[11:39] Strife Onizuka: both VEHICLE_TYPE_NONE and none have the same value, so yes
[11:39] Strife Onizuka: *FALSE
[11:39] Kitto Flora: Is flying-static Avatar spuriously sinking supposed to be fixed? (Its not in Beta).
[11:39] Andrew Linden: Not fixed yet Kitto.
[11:40] Creem Pye: has their been any word from the viewer team as to when they'll remove the 768m limitation for editing prims?
[11:40] Creem Pye: *there
[11:40] Kitto Flora: OK. Thats a minor one.
[11:40] Andrew Linden: Not that I know of... Simon that bug didn't get fixed did it?
[11:40] Nber Medici checks to see if her chat line works... no response to last question.
[11:41] Andrew Linden: Creem, I'll ask about the build limit change. Don't know.
[11:41] Creem Pye: thanks
[11:41] Andrew Linden: Rumor is that it is in the next Release Candidate Client
[11:41] Bato Brendel: its been imported into a internal jira nber. DEV-13264
[11:41] Gaius Goodliffe: Nber: Sidewinder suggested an internal meeting on the subject before he left.
[11:42] Nber Medici: Was not clear to me that ... his comment was in regard to that JIRA
[11:42] Nber Medici: thanks
[11:42] Andrew Linden: Repeat the question Nber... was that about the 2006 jira something or other?
[11:42] Simon Linden: The flying/drifting avatar bug hasn't been fixed yet. I realize it's annoying, but we're giving broken stuff higher priority
[11:42] Nber Medici: I can't tell what the status of the Linden response is to the 2006 JIRA. Can I get a read on that?
[11:42] Nber Medici: yes the 2006 JIRA
[11:42] Andrew Linden: I'm not sure I understand that question. Perhaps you could rephrase it?
[11:43] Gaius Goodliffe: That's the prim-seam problem. You mentioned the seam welder too, but didn't sound like you were sure of the details.
[11:43] Andrew Linden: What is 2006 jira?
[11:43] Arawn Spitteler was drifting earlier, while standing, but not reproducibly.
[11:43] Strife Onizuka: (i'm going fly, thanks for the discussion, take care all)
[11:43] Nber Medici: Golf is broken... it is the seams thing
[11:43] Kitto Flora: Whats the prognosis on cut tapered cylinders having the wrong collision() event distance getting fixed? Should I try to design around it?
[11:43] Andrew Linden: Ah, perhaps you mean SVC-2000 or something?
[11:43] Bato Brendel: shes asking what is LL's word on this issue.... the prim seam issue.
[11:43] Nber Medici: yes
[11:43] Andrew Linden: I'll read it...
[11:44] Nber Medici: what Bato indicated... that is the JIRA that I am referring to.
[11:44] Nber Medici: Thank you Andrew.
[11:45] Andrew Linden: That bug is going to be hard to fix. I'll have to think about it, and maybe talk to some Havok engineers.
[11:45] Kitto Flora: Which bug?
[11:45] Andrew Linden: There may be some workarounds, not sure.
[11:45] Nber Medici: It is vital for our opeation.... and the workaround is a bitch.
[11:45] Andrew Linden: I'm talking about SVC-2006
[11:45] Nber Medici: will involve a LOT of manual editing of many many prims.
[11:46] Creem Pye: has anybody noticed any changes in setting STATUS_GRAB in child prims?
[11:46] Bato Brendel: Kitto in a nutshell, round small balls when rolling over multiple prims jump up when it gets to the prim edge.
[11:46] Kitto Flora: Bato: I uderstand
[11:46] Andrew Linden: well, some bugs I can say, "Oh, I know how to fix that... just got to get around to it". Other bugs fall into the category of "I don't know how to fix it yet."
[11:46] Les White: sounds like you might be on a dead end street with closed scripts Nber
[11:47] Andrew Linden: The prim seams bug currently falls into the second category.
[11:47] Creem Pye: er STATUS_BLOCK_GRAB
[11:47] Bato Brendel: dont work
[11:47] Andrew Linden: Not that those bugs can't be fixed
[11:47] Bato Brendel: or if it does we need some examples on the wiki of a 'working one'
[11:47] Andrew Linden: however right now I cannot see the path to fixing them, or else research will be necessary.
[11:48] Nber Medici: I have brought MTW to discuss the impact on golf of SVC 2006
[11:48] Andrew Linden: What is the STATUS_BLOCK_GRAB issue?
[11:48] MarkTwain White: it is appropirate for me to jump in here?
[11:48] Creem Pye: well it used to be in H1 that you had to set STATUS_BLOCK_GRAB to TRUE in every prim in the linkset, if you wanted to block grabbing for the object
[11:49] Andrew Linden: welcome MarkTwain
[11:49] MarkTwain White: ty Andrew
[11:49] Creem Pye: but I noticed in H4 that it's ineffective in child prims; so I had to delink the object, set the status in each prim, then relink it
[11:49] Creem Pye: maybe that was just a fluke though
[11:49] Andrew Linden: sounds like a bug Creem
[11:49] Bato Brendel: also Ive tried playing with that as well it never worked for me
[11:50] Bato Brendel: to stop people from 'grabbing my boat' and stopping me dead
[11:50] Creem Pye: maybe in H1 it was every broken... it's supposed to affect the entire object, but it only affected individual prims
[11:50] Les White: that's grabbing in edit, not touch. cant get around it
[11:50] Creem Pye: yeah Bato, that's exactly what I was doing
[11:50] Gaius Goodliffe: Bato: That's a different bug I think.
[11:50] MarkTwain White: may i add a comment on svc-2006
[11:50] Bato Brendel: yeah 'right click grab' not 'left'
[11:50] Andrew Linden: right, GRAB is different from EDIT
[11:50] Creem Pye: ah
[11:50] Andrew Linden: Sure MarkTwain
[11:51] Creem Pye: I'm talking about left-click, though EDIT mode blocking would be nice too
[11:51] MarkTwain White: i got a bit of the chat, and undrestand that it is a tough one...
[11:51] Creem Pye: this ticket is about the H1 behavior, but the behavior hasn't improved in H4
[11:51] Gaius Goodliffe: Edit shouldn't stop dead something you can't edit, but it does. Makes cheating in dogfighting trivially easy.
[11:51] MarkTwain White: and ZATZAi Asutiris has found a work aroound....
[11:51] MarkTwain White: but it involved editing prims...
[11:51] Kitto Flora: There's never been a 3rd party Edit block - which is good & evil
[11:51] MarkTwain White: and the golf greens alone are 1800 prims
[11:51] Les White: (if you block grabbing in edit...field day for griefers i think)
[11:52] MarkTwain White: so the cure is pretty painful
[11:52] MarkTwain White: ZAT can help along with Bato on working with LL to solve this issue
[11:52] Creem Pye: Les, maybe there could only be an exception if another avatar is sitting on the object?
[11:52] Les White: nice Creem. works for me :)
[11:53] Arawn Spitteler: Edit Detection?
[11:53] Andrew Linden: Well, I'm willing to work on the prim-seam bump problem. I was just saying that at this moment I don't know how to solve it in a general way.
[11:53] Bato Brendel: allready is one for 'detection'
[11:53] Andrew Linden: Currently there is magic for reducing the problem for vehicles, and even that doesn't work for all vehicles.
[11:53] Gaius Goodliffe: We could also remove LSL entirely, but honestly, breaking the way things work because it helps deal with griefing has never seemed like a good idea to me. The edit thing should be fixed, your "cure" is worse than the disease in this case.
[11:53] Andrew Linden: One solution to solving the prim seam problem may be to move to a different physics engine.
[11:54] Les White: the pain
[11:54] MarkTwain White: understand Andrew. I think ZAT has already spoken to Sidewinder on this,and i know that he and Bato have been working on this at Holly Kai Golf course a lot...
[11:54] Andrew Linden: Another solution may be to get the Havok engineers in and see if they have any ideas.
[11:54] MarkTwain White: just need your ear and Sidewinders, because we can do the detective work
[11:54] Andrew Linden: The prim seam problem is fundamental to Havok, unfortunately... a quirk of their engine design.
[11:55] Bato Brendel: okay Andrew my question is then why now in H4 not H1 I think it would be a matter of looking at the edge det code and then seeing whats majorly changed in it
[11:55] Rex Cronon: somebody said something about "weding"?
[11:55] Nber Medici: welding I think
[11:55] Rex Cronon: welding*
[11:55] Andrew Linden: welding...
[11:56] Gaius Goodliffe: welding, kinda like wedding. ;)
[11:56] Rex Cronon: sorry. typo
[11:56] Andrew Linden: there is some class structure in the Havok API that can be used to "weld" two seams together and fix the seam bump
[11:57] Andrew Linden: however my limited understanding of it was that it was not something that you would do for ALL content... to expensive
[11:57] Andrew Linden: but you might use it for very specific stuff
[11:57] Bato Brendel: now that sounds like a nice feature to add support for. expensive as in how? resources? or coding it in?
[11:57] Les White: i wonder if the balls use some kinda hover and now the larger collision tolerance makes an issue
[11:57] Andrew Linden: Most games can very carefully tweak their content in such ways
[11:57] Les White: sure did for my stuff
[11:58] Nber Medici: We cannot tweak since the scripter is MIA
[11:58] MarkTwain White: one of our problems is the code is orphaned
[11:58] Andrew Linden: however, SL is much more open in the possibility space -- a general solution would have to be found
[11:58] Nber Medici: hehehe
[11:58] Andrew Linden: Yes MarkTwain that had been mentioned.
[11:58] Les White: what are the ethics of re-opening the scripts with linden superpowers i wonder :)
[11:58] Bato Brendel: very good question Les
[11:59] Andrew Linden: There is no policy for doing that. I doubt there will be.
[11:59] Kitto Flora: Who was the scripter?
[11:59] Gaius Goodliffe: Les: It's probably illegal, regardless of the ethical questions...
[11:59] Bato Brendel: Kanker Greenacre
[11:59] Andrew Linden: Er.. I mean no way to open scripts for others.
[11:59] Bato Brendel: same make of the Tako boat
[11:59] Kitto Flora: Oh...
[11:59] MarkTwain White: i am making every effort to track down the scripter, but no success yet, he said he was considering making it open code before he disappeared
[11:59] Bato Brendel: yeah not a 'noname' guy hes pretty well known 8(
[12:00] Gaius Goodliffe: Kanker is MIA? That's a shame...
[12:00] Kitto Flora: YEah... but he may have joined the ranks of the abandoned.
[12:00] MarkTwain White: kanker has not been here since last september
[12:00] Rex Cronon: the ball is scrpted?
[12:00] MarkTwain White: and does not respond to RL inquries
[12:00] Bato Brendel: lemme look I can probally rez one up
[12:00] MarkTwain White: yes it is
[12:01] Andrew Linden: This back in from the "raise the 768 edit limits" ... it will be in the 1.20.0 SL client or higher. So whenever that is releases.
[12:01] Andrew Linden: released
[12:01] : Fastman by Thomaz Wright
[12:01] Creem Pye: ah I see. Hopefully somebody will release a patched client with that limitation removed before then =)
[12:01] MarkTwain White: Bato is playing games with us. AHAHHA
[12:01] Gaius Goodliffe: No modify... yikes.
[12:01] Bato Brendel: well they wanted to see so Im showing...
[12:02] Bato Brendel: yep completely no mod 8(
[12:02] Nber Medici: thanks Bato
[12:02] Gaius Goodliffe: If it was modify, there are workarounds you could try just by dropping in another script, even if you can't edit the first.
[12:02] Rex Cronon: the ball, could be reverse engineered
[12:02] Bato Brendel: well that and as you seen it when physical first
[12:02] Gaius Goodliffe never, ever, ever buys anything no-mod.
[12:02] Bato Brendel: but when it dropped it turns off
[12:02] Nber Medici: Gaius that is NOT the issue here.
[12:03] Gaius Goodliffe: /me/unless it's really cheap.
[12:03] Bato Brendel: now what turns it back on again, is a whole another problem
[12:03] MarkTwain White: this was not bought Gaius, this was programmed for Holly Kai Golf course
[12:03] Gaius Goodliffe nods.
[12:03] Andrew Linden: it uses llApplyImpulse() to actually move
[12:04] Andrew Linden: the bump-collision with the prim seams is the physics engine in action.
[12:04] Creem Pye: that seems like a pretty good method for a golf ball
[12:04] Bato Brendel tries to peek at Andrews screen for a minute
[12:04] Gaius Goodliffe: hehe
[12:05] MarkTwain White: no kidding Bato. i am using telepathy myself
[12:05] Andrew Linden: single prim uniform sphere
[12:05] Bato Brendel nods
[12:05] Andrew Linden: so it should be using the implicit sphere shape for collisions
[12:05] Arawn Spitteler wonders where Hermosa Beach is: This sounds like precisely his reason for not returning to SL, too much like work...
[12:05] Gaius Goodliffe imagines Andrew looking at a Matrix-like rain of text.
[12:05] MarkTwain White: long story
[12:06] Andrew Linden: ack. I'm going to have to go. I've got way too many bugs to fix
[12:06] Nber Medici: ah
[12:06] Nber Medici: :)
[12:06] Les White: thanks for your time Andrew. interesting meeting
[12:06] Creem Pye: thanks for your time
[12:06] MarkTwain White: well touch our golf ball and do a magic fix before you leave
[12:06] Scripting Student goes with mail instead of asking here then based on the response to his IM
[12:06] Bato Brendel: okay andrew lemme ask this I just turned on BBox's and Im not seeing any round ones is there any way for us to see the physics implicit shape for collisions
[12:06] MarkTwain White: hehe
[12:06] Les White: thanks Simon
[12:06] Andrew Linden: no, the visible shape is not necessarily the collision shape
[12:07] Andrew Linden: however, sometimes it is
[12:07] Arawn Spitteler wonders what his head was jsut ding
[12:07] Bato Brendel: yeah Im not looking at its visible shape Im looking at the Yellow and Orange BBox
[12:07] Kitto Flora: You'd think LL would hire a few more bug fixers, would't you?
[12:07] Andrew Linden: there is no way currently to view the exact collision shape
[12:07] Bato Brendel: k so what Im seeing isnt the collision one interesting...
[12:07] MarkTwain White: BBoxes dont tell that?
[12:08] Andrew Linden: uniform spheres ALWAYS use the implicit sphere shape for collisions because that is the cheapest shape available
[12:08] Bato Brendel nods
[12:08] Andrew Linden: convex, nonsheared boxes ALWAYS use boxes because that is the second cheapest shape
[12:08] Simon Linden: Bato - no, there's no easy way to see that. We have a debugging tool that can, but it takes a special build of the sim to get a visual peak at what Havok is using
[12:08] Bato Brendel: ah okay
[12:09] Creem Pye: strange, i'd expect cylinders to be cheaper than boxes...
[12:09] Gaius Goodliffe: Creem: Hmmph. I'
[12:09] Andrew Linden: most concave prims use the lowest level of visual detail shape
[12:09] Bato Brendel wtb a hack piece of code to turn that on as I'll be happy enough to lag and deal with the overload of rendering 8)
[12:09] Gaius Goodliffe: d have expected boxes to be cheaper than spheres.
[12:09] Simon Linden: Assuming that's a sphere without any cuts, dimples, etc, it should be a sphere internally
[12:09] Andrew Linden: spheres are cheapest for memory storage, and also for collision tests
[12:10] Creem Pye: if you think about it, spheres just have 1 collision "region" whereas boxes would have... 22 or so?
[12:10] Bato Brendel: wow I would think a 6 surface box would have been compared to a vertexed round one
[12:10] Andrew Linden: spheres are just a point and a radius
[12:10] Bato Brendel: ah gotcha
[12:10] Bato Brendel: slick thinking that way
[12:10] Simon Linden: I have to run ... bye everyone, see you next time
[12:10] Les White: see you
[12:11] Andrew Linden: thanks for coming
[12:11] Creem Pye: have a good day everybody
[12:11] Rex Cronon: bye simon
[12:11] MarkTwain White: thank you andrew
[12:11] Gaius Goodliffe: Creem: Well, it's easier to tell if someone is inside a box than a sphere usually. Actually, not quite as true when it's rotated. Nevermind... :p
[12:11] Bato Brendel: no thanks you for having us Andrew
[12:11] Rex Cronon: bye andrew
[12:11] Nber Medici: thanks :)
[12:11] Kitto Flora: Byeby
[12:11] Nber Medici: /bye