Simulator User Group/Transcripts/2012.03.30

From Second Life Wiki
Jump to navigation Jump to search

Simulator_User_Group

Prev 2012.03.27 Next 2012.04.03

List of Speakers

Andrew Linden AquilaDellaNotte Kondor Ardy Lay
Ashiri Sands Chieron Tenk Christy Mynx
Draconis Neurocam FadeOut Razorfen Falcon Linden
Fancy Detector Fenix Eldritch Flip Idlemind
MartinRJ Fayray Motor Loon Reno Parkin
Rex Cronon Simon Kline Sniper Siemens
Squirrel Wood TankMaster Finesmith Vincent Nacon

Transcript

[15:56] MartinRJ Fayray: Hello

[15:56] Motor Loon: ey martin

[15:56] Ardy Lay: Boogie with the babes? MOTD is off the wall.

[15:56] Chieron Tenk: heyas

[15:57] FadeOut Razorfen: oi, firestorm betatester, i just wasted 30 minutes of my life on figuring out that in build menu link number display of prim is inverted, gogo take care

[15:58] Sniper Siemens: -)

[16:00] Meeter: Welcome to the Server User Group

[16:01] Flip Idlemind: One assumes Lindens will be here

[16:01] Ashiri Sands: One hopes

[16:01] Vincent Nacon: yup, they tend to be a minute or 5 late

[16:01] Motor Loon: Late lindens...

[16:01] Ashiri Sands: ssuming is not always prudent

[16:01] Vincent Nacon: they're only humans

[16:02] Motor Loon: well,... semihumans atleast

[16:02] Vincent Nacon: -_-;

[16:02] Draconis Neurocam: i saw simon log in twice in the exorbitant time i have been sitting here

[16:02] Vincent Nacon: they're human...

[16:02] Chieron Tenk: so far it's indeed a user meeting

[16:02] Vincent Nacon: ahh sounds like Simon is doing quick chores

[16:02] Fenix Eldritch: first one I've been able to attend in a loooong time

[16:02] Vincent Nacon: I wonder why is that

[16:03] Chieron Tenk: I feel less human when sitting 10cm above the chair, tho.

[16:04] Vincent Nacon: maybe because you're ...some type of cat?

[16:04] Vincent Nacon: start with an S I think

[16:04] Rex Cronon: gretings all

[16:04] Vincent Nacon: heya Rex

[16:04] MartinRJ Fayray: Hello Rex

[16:04] Motor Loon: ey rex

[16:05] Rex Cronon: hi vincent, martinrj, motor

[16:05] Rex Cronon: no lindens?

[16:05] Vincent Nacon: they're just running a bit late

[16:05] Chieron Tenk: hi rex

[16:05] Rex Cronon: hi chieron

[16:06] Vincent Nacon: could be a good news though

[16:06] Vincent Nacon: meaning they're doing something

[16:06] Motor Loon: could also mean the serverroom is on fire...

[16:06] Chieron Tenk: serval, vincent.. no, it's because of the sittarget being slightly too high

[16:06] Rex Cronon: they r working hard to keep this place going

[16:06] Rex Cronon: AFK...............................

[16:06] Rex Cronon: i will be afk

[16:07] Chieron Tenk: we're alle levitating a bit, heh.

[16:07] Simon Kline: hi all, no lindens today? :D

[16:08] Vincent Nacon: they're just running a bit late

[16:08] Fancy Detector: Andrew Linden has arrived!

[16:09] Fenix Eldritch: Hullo

[16:09] Sniper Siemens: Hi Andrew

[16:09] Simon Kline: thanks vincent :D

[16:09] Andrew Linden: Hello

[16:09] MartinRJ Fayray: Hello Andrew!

[16:09] Vincent Nacon: Chieron, I think you're having viewer-side problem with your sit placement

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

[16:10] Vincent Nacon: because I saw you sitting on the chair just fine

[16:10] AquilaDellaNotte Kondor: hello everybody

[16:10] Andrew Linden: I was tracking down a crash bug in the pathfinding project and fogot the time.

[16:10] Ashiri Sands: I see levitation too

[16:10] Vincent Nacon: hmm

[16:10] Vincent Nacon: oh well

[16:10] Vincent Nacon: welcome Andrew, Simon on the way?

[16:11] Andrew Linden: Hrm... Simon appears to be out of the office.

[16:11] Andrew Linden: Which probably explains why he didn't ping me when I was late.

[16:11] Rex Cronon: back

[16:11] Christy Mynx: wb

[16:11] Vincent Nacon: muhaha!

[16:12] Rex Cronon: hello andrew

[16:12] Andrew Linden: Ah yes, he's on vacation.

[16:12] Chieron Tenk: no excuse for not pinging.

[16:12] Vincent Nacon: oh yeah

[16:12] Andrew Linden: I think he was out last Friday too, but he warned me about that one.

[16:12] Vincent Nacon: he did mention that last meeting

[16:12] Andrew Linden: Anyway... news.

[16:13] Andrew Linden: We've opened up the Pathfinding beta channel, whch is the project I've been working on for the last several months

[16:13] Andrew Linden: we had a bunch of crashes and performance issues

[16:13] Andrew Linden: so Falcon and I have been scrambling to fix various problems.

[16:13] Flip Idlemind: Good (that you're catching them now, I mean)

[16:14] Andrew Linden: This is just a micro channel of regions on the grid, for volunteers.

[16:14] Andrew Linden: So that project has been what has keeping me from doing more maintenance stuff these days.

[16:15] Andrew Linden: But it looks like it is wrapping up soon (a few weeks?) so I should be able to get to do more fixing rather than features.

[16:15] Andrew Linden: BTW, I finally was able to update the LSL calls in the pathfinding viewer's tooltips

[16:16] Andrew Linden: so those should be merged into the main viewer project eventually

[16:16] Andrew Linden: I also updated the keywords.ini file, but not quite to my satisfaction.

[16:16] Vincent Nacon: but good enough to go?

[16:16] Andrew Linden: I had to leave a lot of description fields blank because I didn't have enough time to research and come up with copy.

[16:17] Andrew Linden: I've got some copy from Ima Mechanique, but some of it is too verbose for what I wanted

[16:17] Draconis Neurocam: awesome that you got it done to any extent

[16:17] Andrew Linden: and I didn't have time to trim it down to more terse sentences.

[16:17] Vincent Nacon: quick! someone smack Ima!

[16:17] Vincent Nacon: oh not here...

[16:18] Andrew Linden: Ima is working on a more advanced syntax hilight file for an external LSL editor

[16:18] Rex Cronon: and this new and improved keywords.ini file can be found where?

[16:18] Andrew Linden: but right now the keywords (real keywords like "for" and "while") and the LSL calls (llSay()) are in separate files

[16:19] Andrew Linden: they should probably be put into one file, and then served via some http request by each region

[16:19] Andrew Linden: and the viewer should be able to load a new syntax file on the fly

[16:19] Rex Cronon: maybe an index file that point to where each individual ini file is located?

[16:19] Andrew Linden: if we ever get to that state then we may be able to use the same syntax file as Ima's editor

[16:20] Flip Idlemind: How will that handle functions that exist but haven't been released? (Not that I know any like that <.<)

[16:20] Andrew Linden: or come to some sort of agreement on a standard format

[16:20] Fenix Eldritch: including block comments?

[16:20] Draconis Neurocam: block comments work in the default viewer now fenix

[16:21] Fenix Eldritch: oh? ha ha, nevermind

[16:21] Andrew Linden: well Flip, ideally if a new LSL call existed on some region that was not yet released (such as an experiment on aditi) then you could download that region's syntax from there when you visit

[16:21] Rex Cronon: block comments? u mean like /* comments */

[16:21] Draconis Neurocam: yeah

[16:21] Rex Cronon: lsl supports that?

[16:21] Draconis Neurocam: tes

[16:21] Draconis Neurocam: yes*

[16:22] Andrew Linden: That is all the news that I've got.

[16:22] Andrew Linden: Table is open.

[16:23] Fenix Eldritch: I have a question RE pathfinding and doors...

[16:23] Andrew Linden: ah yes, the house doors problem

[16:23] Rex Cronon: hmmm. nice:)

[16:23] Fenix Eldritch: So the current implementation basically only allows for the "Movable Obstacles" to be applied to "doors". Meaning they may open, but as far as the character is concerned, it's something blocking the doorway and it won't bother investigating for a way to open it, correct?

[16:23] Andrew Linden: yes, the characters won't be smart enough to open doors

[16:24] Fenix Eldritch: I know it's too late in the game for more features, but my I suggest a thought?

[16:24] Andrew Linden: that is, there won't be support for door intelligence in the characters (at leasts not in the first version)

[16:24] Rex Cronon: when it bumps into a door it should shout "open sesame":)

[16:24] Andrew Linden: go ahead Fenix

[16:24] Vincent Nacon: ew... spammy door

[16:24] FadeOut Razorfen: there are several workarounds for that tho

[16:24] Fenix Eldritch: Essentially, I'm envisioning this being applied to a door which when closed, the character would still "know" it can get through it and plots its path as if it were open. When it reaches the door and senses it cannot progress, it would make use of LSL logic to interact with the (scripted) object.

[16:25] Fenix Eldritch: Sorry, I meant to say before having a new type of classification of object

[16:25] Fenix Eldritch: one which the character does not try to avoid, but rather stops to inspect when/if it blocks their path.

[16:25] Fenix Eldritch: This could trigger a new path_update event... something like PU_INSPECTING which scripters could use to trigger their interaction logic.

[16:26] Vincent Nacon: yeah, sounds like you have point there

[16:26] Andrew Linden: The havok AI engine has a feature for connecting two distinct navigation meshes that are connected by special routs/events

[16:26] FadeOut Razorfen: actually... path is pre-computed

[16:26] Andrew Linden: I think they are called "user edges" or something like that

[16:27] Fenix Eldritch: Ah my thought is that the new obstacle typ would not register as a true obstacle on the nav mesh

[16:27] FadeOut Razorfen: ohcool

[16:27] Vincent Nacon: I also wondered about character ability to jump, some setting to limit how far they can jump

[16:27] Andrew Linden: so one could define a way for an AI character to know that it can hop from one mesh to another

[16:27] Vincent Nacon: and how high

[16:27] Andrew Linden: for example, via a moving platform that had its own mesh

[16:27] Andrew Linden: or via a door that could somehow be opened

[16:28] Andrew Linden: however, we're not tapping into that feature yet for the first implementation

[16:28] Vincent Nacon: as for the door, they'd just have to be aware of door's location and use Nav function to get the door

[16:28] Fenix Eldritch: right, just a thought

[16:28] Andrew Linden: we do want to investigate possible feature enhancements to pathfinding later, using stuff like that

[16:28] Andrew Linden: however... one "problem" with doing that

[16:29] Fenix Eldritch: what to do if it can't get past it?

[16:29] Vincent Nacon: permission abuse?

[16:29] Andrew Linden: is that there is a lot of UI that needs to be made, and polished, and remade

[16:29] Fenix Eldritch: ah

[16:29] Vincent Nacon: oh

[16:29] Andrew Linden: the server-side work might not be too hard

[16:29] Fancy Detector: Falcon Linden has arrived!

[16:30] Andrew Linden: but providing the tools for content creators to set up complex transitions between distinct navigation meshes, or across otherwise unnavigable obstacles

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

[16:30] Andrew Linden: would be more work

[16:30] Falcon Linden: new furniture again?

[16:30] Vincent Nacon: heya Falcon, we were just talking about pathfinding, character to know what door it could go through

[16:30] Falcon Linden: Vincent: you mean, how can you build a door that works with characters?

[16:30] Andrew Linden: Falcon could talk more about that (user edges between distinct navigation meshes)

[16:31] Vincent Nacon: pretty much, or the "interaction" part of it

[16:31] FadeOut Razorfen: ohgod ohgod, falcon, did you notice those new jiras ? https://jira.secondlife.com/browse/PATHBUG-65 and https://jira.secondlife.com/browse/PATHBUG-64

[16:31] Andrew Linden: Not so much about how to create a door in pathfinding as it is today, Falcon

[16:31] Andrew Linden: but more about how to make a door such that a character would know how to open it

[16:31] Vincent Nacon: hang on Fade....

[16:31] Falcon Linden: okay, so here's how it would work in a perfect world in the future:

[16:32] Falcon Linden: 1) You set up some trigger volumes in front of and behind the door or you use at_target to determine when you've reached a door.

[16:32] Falcon Linden: Well, actually, I take that back. It definitely depends on the underlying door mechanics

[16:32] Draconis Neurocam: at_target is so underused

[16:32] Falcon Linden: In a perfect world, the door wouldn't be part of the navmesh. Instead, when a character reached an area in front of the door, it would stop and trigger the door to open

[16:33] Fenix Eldritch: Yeah, I had the thought to raise a new path_update event when getting stopped by a special door obstacle

[16:33] Falcon Linden: then it would continue pathing through and on its way out, trigger the door to close

[16:33] Falcon Linden: but because the door wasn't in the navmesh, the AI query would just be "get me to my goal" not "get my to the door and figure out what happens next"

[16:33] Vincent Nacon: there's two ways I see it... set door type for character to understand what they are. or script the charadcter to set new Nav point base on the location of the door.

[16:33] Falcon Linden: That's assuming it's a door that a character can open itself

[16:34] Falcon Linden: If you want to do it today, you use the second option from what Vincent just suggested. You:

[16:34] Falcon Linden: 1) Have you character navigate to in front of the door

[16:34] Falcon Linden: 2) Have the character send a message to the door triggering it to open via llSetKeyframedMotion (the only way to have a door that works with pathfinding at all)

[16:34] Vincent Nacon: yeah and have the door work with the character to open for it

[16:34] Falcon Linden: 3) Path through

[16:34] Falcon Linden: yep

[16:34] FadeOut Razorfen: oh thats why zombies in L4D2 start breaking random doors they don't actually need to achieve their goal lol

[16:34] Vincent Nacon: muhaha!

[16:34] Rex Cronon: when u say "character" u mean bot or ave?

[16:35] Falcon Linden: I mean a pathfinding scripted object

[16:35] Vincent Nacon: bot

[16:35] Rex Cronon: ok

[16:35] Falcon Linden: shoot, I left something at the front desk, brb (2secs)

[16:35] Andrew Linden: we mean "AI character" = object bot using the pathfinding stuff via LSL script

[16:35] Vincent Nacon: but of course... we're still assuming we can use avatar for bot

[16:35] Vincent Nacon: in the future

[16:35] Falcon Linden: back, sorry

[16:36] Andrew Linden: Yes, that would be a grand future, so LL would like to get that done too.

[16:36] Rex Cronon: 2 me character is a little confusing. makes me think of avatars

[16:36] Falcon Linden: any term will be confusing

[16:36] Vincent Nacon: anyway... what about AI dealing with edges to climb, jump, or even "hang by hands crawl" ?

[16:36] Chieron Tenk: llNPC would look weird, tho

[16:36] Andrew Linden: Yeah, I didn't like it at first either. I'm not so bothered anymore

[16:37] Andrew Linden: but I can't think of a better term

[16:37] Falcon Linden: NPC would probably have been best, but it's problematic too.

[16:37] Falcon Linden: Vincent: custom scripting for that

[16:37] Reno Parkin: Hello, can i switch to a new question

[16:37] Andrew Linden: Sure Reno, go ahead.

[16:37] Reno Parkin: i wanted to know how the LSL script indexes multiple object collusions at same time (with in a span of 0.13 seconds). is it like the first objects that collided will have the index of 0? I wanted to know this as i am creating a car racing project and there can be more than one care getting collided with the finish line with in span of 0.13 seconds.

[16:37] Vincent Nacon: no I mean... in part of the pathfinding

[16:37] Falcon Linden: vincent: in principle, we could add stuff like that in the future with Havok's tools, but more advanced features become more and more tricky to make work right "in general"

[16:37] Falcon Linden: vincent: trigger volumes/at_target that start/end points of those sections.

[16:38] Vincent Nacon: so it could know it can get to its location by doing those

[16:38] Vincent Nacon: hmm

[16:38] Falcon Linden: with communication to/from the character

[16:38] Falcon Linden: we don't have tools right now to support the "right" solution of custom edges between locations

[16:38] Vincent Nacon: right

[16:38] Falcon Linden: (so-called "user edges"). We toyed with the idea early on but the tools quickly became unwieldy

[16:39] Vincent Nacon: well what do you mean by unwieldy? too advanaged?

[16:39] Andrew Linden: Hrm... Reno, I think the collision events really do queue up in order

[16:39] Falcon Linden: Reno: It's unspecified behavior

[16:39] Falcon Linden: Reno: and not guaranteed to be the same in future releases

[16:39] Vincent Nacon: advanced*

[16:39] Falcon Linden: vincent: too hard to make usable tools to create them, too hard to write code that ensures that they are valid, difficult interaction model with scripts, etc.

[16:39] Andrew Linden: but the events may have actually happened at different frames... it might take a few frames for the script to get a turn to handle them, I think.

[16:39] Vincent Nacon: hm ok

[16:39] Falcon Linden: basically, they work well for a game studio that has full control of tools and pipelines, but not well for us

[16:40] Vincent Nacon: instead of tools, think it could be done by LSL with some list?

[16:40] Vincent Nacon: assign by vector in list I mean

[16:40] Falcon Linden: we thought of that too and decided it was too hard to right the code that ensures they are legal and too hard to deal with managing these things

[16:40] Vincent Nacon: ah maybe later in future then

[16:41] Falcon Linden: how do you remove them? Change them? Find out which ones exist? We came up with solutions but it became clear that the few users who would likely choose to use it weren't worth the effort

[16:41] Reno Parkin: ok, can u tell me where exactly i can find information about the official order for indexing of collusion event

[16:41] Falcon Linden: but yes, in the glorious future, particularly if we ever get C# or other "real" language..

[16:41] Falcon Linden: Reno: there isn't one

[16:41] Falcon Linden: Reno: And even if we tell you how it works now (which I'm not sure we know for certain without looking at the code), it's subject to change

[16:42] Falcon Linden: it's one of those "micro" features we can't easily guarantee will work with future physics changes, etc.

[16:42] FadeOut Razorfen: right now it does sort by order, 0 is first collision... but what falcon said

[16:42] Falcon Linden: do NOT rely on that. Maybe I'll randomize it just so you don't :P

[16:43] FadeOut Razorfen: :3

[16:43] Falcon Linden: By the way, speaking of things I've asked people not to do that they keep doing: You no longer need to call llDeleteCharacter before llCreateCharacter

[16:43] Falcon Linden: see the wiki

[16:43] Falcon Linden: If I keep seeing it, I'm going to add a script penalty for calling delete :P

[16:43] Vincent Nacon: overided?

[16:43] Vincent Nacon: overrides*

[16:43] Falcon Linden: llCreateCharacter will reset all params to their defaults other than size/orientation and it will apply the provided settings

[16:44] Squirrel Wood: just make it an empty function call instead that does nothing ?

[16:44] Vincent Nacon: alrighty, sounds good

[16:44] Falcon Linden: it is low impact to the server and easy to use. llDeleteCharacter, OTOH, is expensive

[16:44] Falcon Linden: (but sometimes necessary if you really want to become a physical object or something again)

[16:44] FadeOut Razorfen: oh, k, sorry :D

[16:44] Vincent Nacon: hmm... dare I ask about character's flying ability?

[16:44] Fenix Eldritch: Why is is so expensive?

[16:44] Fenix Eldritch: *it

[16:44] Falcon Linden: Vincent: On my list of priorities for "if we get time" or "in my spare time"--it's underneath creating a flocking behavior, though

[16:45] Vincent Nacon: ah interesting

[16:45] Falcon Linden: Fenix: It forces us to rebuild the physics shape for the object and add it to the world. That's an expensive operation

[16:45] Fenix Eldritch: ah

[16:45] Flip Idlemind: So, this JIRA is old-ish, but it's been on my mind: SVC-3811

[16:45] Flip Idlemind: Im curious what people think of my comment: "While we're at it, could we get rid of the whole "No room to sit here" thing all together? What purpose does that serve really?"

[16:45] JIRA-helper: http://jira.secondlife.com/browse/SVC-3811

[#SVC-3811] Sit disabled on prims that are: inside a prim's bounding box (eg, phantom prim).

[16:45] Ashiri Sands: Flocking behaviour would be nice

[16:45] Falcon Linden: that's a bug inside how we find seat positions, but it's become a feature

[16:45] Falcon Linden: so it's unlikely to get fixed

[16:46] Rex Cronon: isn't there a fix for that?

[16:46] FadeOut Razorfen: falcon, did you notice those new jiras ? https://jira.secondlife.com/browse/PATHBUG-65 and https://jira.secondlife.com/browse/PATHBUG-64

[16:46] Vincent Nacon: what about setting the "apple" to none physic?

[16:46] FadeOut Razorfen: some really important stuff thar

[16:46] Falcon Linden: Fadeout: not till just now, but yes, those look like significant bugs.

[16:46] Vincent Nacon: I don't mean phantom

[16:46] Flip Idlemind: I can't set other peoples' griefer objects to no-physical-shape

[16:46] Falcon Linden: Right now I'm dealing with more significant bugs, though, like a few crashes, some very gnarly performance under rare circumstances, and vehicles issues

[16:47] FadeOut Razorfen: yeah, i was crashing hippotropolis like 10 times a day

[16:47] FadeOut Razorfen: while testing stuff

[16:47] Falcon Linden: Yeah, that was my fault. Mismatch in some code. Whoops. The super-easy crasher should be fixed now

[16:47] Andrew Linden: today or yesterday FadeOut?

[16:47] FadeOut Razorfen: yesterday i think

[16:47] Falcon Linden: (basically and character colliding with anything that wasn't convex or the ground would crash)

[16:48] Falcon Linden: Right now, the biggest trouble spot for regions on the beta appears to be a few "badly behaved" objects that are frequently changing their shape

[16:48] Falcon Linden: Although I'm working on some changes to improve the behavior there, those objcts should really be set phantom and have a physics box places around them or removed entirely

[16:49] Falcon Linden: because they're evil. Sadly, they're also the only way to do animation right now. But that doesn't make them less evil.

[16:49] Falcon Linden: One good solution: Link all the pieces to an invisible box and make the visible pieces have no physics shape

[16:49] FadeOut Razorfen: btw, falcon do you aware of small bug that character continues moving on very slow speed when it's stopped moving by script

[16:50] FadeOut Razorfen: last velocity after stop presists

[16:50] FadeOut Razorfen: on llPursue at least

[16:50] Andrew Linden: (btw, eventually that "no physics shape" trick for low-impact visual animation will also work for objects that contribute to the navmesh calculations)

[16:50] Falcon Linden: fadeout: no...there is a minimum speed of 0.02m/s, but that shouldn't apply when the character behavior ahs been terminated via CHARACTER_CMD_STOP

[16:50] Andrew Linden: (but I've yet to get around to doing that)

[16:51] Falcon Linden: fadeout: also, a character on a slope may slide down slowly...I'm not sure.

[16:51] Falcon Linden: feel free to file a bug report.

[16:51] FadeOut Razorfen: okay

[16:51] FadeOut Razorfen: i'l have to capture a movie for that one

[16:51] FadeOut Razorfen: fffffffffffff

[16:52] Falcon Linden: ok. Be sure to include a simple script repro to go with it

[16:52] Falcon Linden: movies alone don't help us much

[16:52] Fenix Eldritch: Falcon and/or Andrew, where would be a good place to direct some more questions I have on the earlier topic I brought up? Forums, Jira, email?

[16:52] FadeOut Razorfen: sure

[16:53] Andrew Linden: Fenix, summarize that topic for me? I can't remember everything that scrolled by today.

[16:53] Falcon Linden: What topic?

[16:53] AquilaDellaNotte Kondor: hi guys

[16:53] AquilaDellaNotte Kondor: see u

[16:53] Fenix Eldritch: The lofty future feature of working with doors

[16:54] Falcon Linden: Jirafy :)

[16:54] Fenix Eldritch: .. er along with pathfinding

[16:54] Fenix Eldritch: ok

[16:54] Falcon Linden: preferably with a description of how you'd like it to work

[16:54] Falcon Linden: I'll at least look at it and try to have a discussion on there (time permitting) that can inform future work

[16:54] Flip Idlemind: So I'm curious, how is not being able to sit on something a feature?

[16:55] Falcon Linden: People have used big phantom prims to keep folks from sitting in a certain area

[16:55] Falcon Linden: like clubs

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

[16:55] Falcon Linden: I remember fixing that problem once upon a time and peope got very angry

[16:55] Fenix Eldritch: cool thanks! I thought I had a simple idea for it, but maybe I still don't grasp how it really works

[16:55] MartinRJ Fayray: whoa

[16:56] Falcon Linden: fenix: but like I siad, it's a bit of a reach. Also, I'd suggest doing some research first into how other games do it :)

[16:57] Falcon Linden: and be sure to play with the pathfinding code as it is (including view visualization tools) to understand how it currently works

[16:57] Fenix Eldritch: right

[16:57] Falcon Linden: Also, who wants to volunteer to build a house with working door example?

[16:57] Falcon Linden: Doesn't have to be character-interactive, but it should allow a user to have a character follow it through the door

[16:57] Rex Cronon: what do u mean falcon

[16:57] Falcon Linden: I have a pet following me with llPursue

[16:57] Falcon Linden: I go to the door

[16:57] Falcon Linden: I open the door

[16:57] Falcon Linden: I walk through and the pet follows

[16:57] Falcon Linden: I close the door

[16:57] Falcon Linden: hint: it'll involve the door using llSetKeyframedMotion and being a separate linkset from the house

[16:58] Falcon Linden: so if the example includes code for how to keep the door position synced with the house position, you get bonus points!

[16:58] Rex Cronon: u don't need a house. something like a simple maze would work too

[16:58] FadeOut Razorfen: really simple rotational math

[16:58] Falcon Linden: Yes, but people pathfinding enabled houses will probably be more common than pathfinding enabled mazes with moving doorways :)

[16:58] Andrew Linden: The door can't be a piece linked to the house (assuming the house actually has a floor (isn't a dirt floor))

[16:59] Andrew Linden: because the house is an object that is included in the calculation of the navigation mesh

[16:59] Falcon Linden: The tricky part will really be the coordination with the house So when you move the house the door moves, rezzing the house rezzes the door from inside it, etc

[16:59] Falcon Linden: I know there are standard tricks for doing that in the building/scripting community

[16:59] Andrew Linden: and such objects can't change shape all the time

[16:59] FadeOut Razorfen: not_at_target

[16:59] Falcon Linden: or ever.

[16:59] FadeOut Razorfen: not_at_rot_target

[16:59] Andrew Linden: (the cost of computing the navigation mesh is expensive, and is not something that is done real-time in most games)

[16:59] Andrew Linden: but we have to be able to do it realtime

[17:00] Falcon Linden: The door also needs to use llSetKeyframedMotion or various throttles/safety mechanisms may keep the changes in door position from being reflected to the navmesh

[17:00] Meeter: Thank you for coming to the Server User Group

[17:00] Andrew Linden: so we threaded it, and it takes tens fo seconds to complete

[17:00] Rex Cronon: i don't know. usually doors come attached to the house

[17:00] FadeOut Razorfen: okay how do i contact you after i make it ?

[17:00] Falcon Linden: but not always.

[17:01] Vincent Nacon: thanks for coming out... time for my dinner!

[17:01] Ardy Lay redesigns a door script.

[17:01] Falcon Linden: fadeout: give me a copy on Agni and send me an IM (not a notecard, I don't read those)

[17:01] FadeOut Razorfen: okays

[17:01] Falcon Linden: thanks folks! Now off to pathfinding work :)

[17:02] Rex Cronon: tc falcon

[17:02] Andrew Linden: If someone sends me the first 8-minutes of the transcripts where you're all wondering why I'm late, I'll include it in the record

[17:02] TankMaster Finesmith: have a grate weekend, anrew, falcon

[17:02] MartinRJ Fayray: goodbye everyone

[17:02] Andrew Linden: thanks for coming

[17:02] Motor Loon: Ill do that andrew

[17:02] MartinRJ Fayray: Goodbye Andrew

[17:02] Rex Cronon: tc andrew

[17:02] Andrew Linden: Thanks Motor

[17:02] Fenix Eldritch: Thanks!

[17:02] Flip Idlemind plants SVC-7743 in everyone's brains

[17:02] Chieron Tenk: bye everyone

[17:02] JIRA-helper: http://jira.secondlife.com/browse/SVC-7743

[#SVC-7743] Better Administrative Control over who can Speak on Voice

[17:02] Rex Cronon: tc everybody

[17:02] Rex Cronon: have a nice day all:)

[17:02] Reno Parkin: thank you all :)

[17:02] Christy Mynx: bye all, take care :)

[17:03] Andrew Linden: excellent

[17:03] Motor Loon: tc all

[17:03] Ashiri Sands thinks of door scripts...



Simulator_User_Group

Prev 2012.03.27 Next 2012.04.03