Simulator User Group/Transcripts/2012.02.10

From Second Life Wiki
Jump to: navigation, search

Simulator_User_Group

Prev 2012.02.07 Next 2012.02.14

List of Speakers

Andrew Linden Arawn Spitteler Ardy Lay
Ash Qin Ayamo Nozaki Davido Chrome
Draconis Neurocam Equino Faulkland IEvangeline
Johan Laurasia Jonathan Yap Kadah Coba
Kallista Destiny Kandi Vendetta Keli Kyrie
Kelly Linden Kuroyu Canis Maggie Darwin
Rex Cronon Sahkolihaa Contepomi Sebastean Steamweaver
Simon Linden Sniper Siemens Squirrel Wood
Vincent Nacon

Transcript

[16:01] Sahkolihaa Contepomi: >____>

[16:01] Ayamo Nozaki: best entrance

[16:01] Vincent Nacon: he broke his table

[16:01] IEvangeline: There we go. 16 scripts

[16:01] Vincent Nacon: or something

[16:01] Sahkolihaa Contepomi: Hey Andrew.

[16:01] Ardy Lay: Wow

[16:01] Ardy Lay: Chart

[16:01] Ardy Lay: That was ugly

[16:01] Johan Laurasia: Looks like a full house today

[16:01] Andrew Linden: Hello

[16:01] Sebastean Steamweaver: Quite!

[16:02] Rex Cronon: hi everybody

[16:02] Vincent Nacon: not really

[16:02] Rex Cronon: hi all

[16:02] Vincent Nacon: full house would have been when all of the angry sailers here

[16:02] Kallista Destiny: Ave Andrew

[16:02] Sahkolihaa Contepomi: Oh let's not go there again.

[16:02] Vincent Nacon: muhaha!

[16:02] Sahkolihaa Contepomi: Hey Simon.

[16:02] IEvangeline: Hi Mr Simon :D

[16:02] Kallista Destiny: occupy linden village

[16:03] Simon Linden: Hello everyone

[16:03] Johan Laurasia: Hey Simon

[16:03] Sniper Siemens: hello Simon

[16:03] Sebastean Steamweaver: Hey Simon, Andrew

[16:03] Kallista Destiny: Hi Simon

[16:03] Ardy Lay: Hello Simon and Andrew

[16:03] Andrew Linden: ok news...

[16:03] Davido Chrome: Hey

[16:03] Rex Cronon: hi

[16:04] Ardy Lay: I think we are pushing maximum nametag density. They are moshing around.

[16:04] Maggie Darwin: Linden viewr crash rates are the suck again. Welcome back to v1.22.

[16:05] Andrew Linden: Hrm... I don't have any news. Simon?

[16:05] Ardy Lay: This? http://wiki.secondlife.com/wiki/Third_Party_Viewer_Directory

[16:05] Simon Linden: Not much ... next week we're looking to promote the code in the release channels to the main one

[16:05] Ardy Lay: Is 3.2.9 worse than that?

[16:05] Rex Cronon: my chat takes 3 minutes to show?

[16:06] Ardy Lay: Rex, get closer

[16:06] Simon Linden: The new ones going into the RCs are centered on the region crossing work ... there isn't any expected improvement yet, but it's the foundation for better performance in the future

[16:06] Rex Cronon: right:)

[16:07] Ardy Lay: Do regions sleep or are they merely daydreaming?

[16:07] Simon Linden: On the viewer crash rate, I know it got pretty bad in January with memory bloat - usage would expand in busy areas until you crashed

[16:07] Simon Linden: I had a hard time lasting through the user group

[16:07] Kallista Destiny: They dream of electric sheep.

[16:07] Maggie Darwin: We call that a "leak".

[16:07] Simon Linden: But the latest beta is much better

[16:07] Davido Chrome: Could you Lindens take a copy of my bike and explain to me why my physics optimisations cause to sink into everything?

[16:07] Davido Chrome: it*

[16:08] Ardy Lay: I am finding I really don't have that crash mode now.

[16:08] Arawn Spitteler: Sinking bike sounds like an attachment, needing an animation

[16:08] Vincent Nacon: hang on david, they're giving news first

[16:08] Simon Linden: That's it for me ... Kelly, do you have anythign?

[16:09] Kuroyu Canis: AYAMOOOO\

[16:09] Rex Cronon: its optimization code is so optimezed that is causing molecular breakdown:)

[16:09] Maggie Darwin: They don't have any news.

[16:09] Arawn Spitteler wishes notices came with expirations.

[16:09] Simon Linden: ok, I guess the floor (no table today) is open

[16:10] Andrew Linden: Davido was wondering about his bike...

[16:10] Vincent Nacon summons a table

[16:10] Ardy Lay: 100% transparent objects were "optimized out" of the rendering pipeline. A change was made to viewer code today to put them back when Highlight Transparent is active.

[16:10] Sahkolihaa Contepomi: I'm running that build, Ardy.

[16:10] Maggie Darwin summons the cops.

[16:11] Jonathan Yap: Oz is on vacatin, so nothing will be going into viewer-development for a while

[16:11] Davido Chrome: Used perfect sphere prims for collision shapes for the wheels.

[16:11] Ardy Lay: In theory

[16:11] Kelly Linden: I don't have any news.

[16:12] Vincent Nacon: I wouldn't suggest sphere for wheels

[16:12] Vincent Nacon: nor sculpty

[16:12] Kallista Destiny: is everyone aware of pink-eye?

[16:12] Vincent Nacon: yes, in RL I do

[16:12] Sahkolihaa Contepomi: Not even 15 minutes in and my viewer crashes: 2012-02-11T00:10:52Z WARNING: LLRenderTarget::allocateDepth: Unable to allocate depth buffer for render target.

2012-02-11T00:10:52Z WARNING: LLRenderTarget::allocate: Failed to allocate depth buffer for render target.

[16:13] Andrew Linden: I don't see anything obviously wrong about the bike. I'd expect it to ride right on top of the ground.

[16:13] Vincent Nacon: yeah I don't see anything wrong other than using sculpty and sphere for physic

[16:13] Davido Chrome: If I hit the sligtest inclination I sink right into it.

[16:13] Andrew Linden: a sphere for wheels is actually a pretty good idea (as far as making for low physics cost)

[16:13] Davido Chrome: No sculpty physics...

[16:14] Vincent Nacon: yeah... David, you're running into the "edge collision" issue

[16:14] Andrew Linden: oh....

[16:14] Vincent Nacon: sphere isn't the best choice for that

[16:14] Rex Cronon: that is normal. the terrain is not that precise

[16:14] Johan Laurasia: what we need are axels

[16:14] Johan Laurasia: axles*

[16:14] Simon Linden: It's fine on flat land?

[16:14] Johan Laurasia: joints

[16:14] Vincent Nacon: either box or cube with flat corners

[16:14] Vincent Nacon: edge*

[16:14] Johan Laurasia: jointed primmys

[16:14] Davido Chrome: Yeah. Flat land works okay. But an incline, I am in the terrain, or the tramp prim.

[16:15] Davido Chrome: ramp*

[16:15] Andrew Linden: Sinking into an edge or step might be caused by some collision-rejection-magic that ground vehicles have

[16:15] Vincent Nacon: David, the issue is part of the Havok's engine, something you'd see in other games with that physic engine

[16:15] Keli Kyrie serves pizza and kool-aid

[16:15] Andrew Linden: which was added to reduce the "bounce at seams in the road" where prims butt up against each other

[16:15] Davido Chrome: The thing is, the problem got progressively worse as I made the bikes physics cost cheaper. Totally counter intuitive...

[16:16] Vincent Nacon: with mesh?

[16:16] Squirrel Wood: the bounding boxes and physics shapes look okay.. though havok tends to simplify shapes under stress

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

[16:16] Davido Chrome: No. By define things as physics shape none.

[16:16] Vincent Nacon: err...

[16:16] Andrew Linden: ok Davido, some info about how that works, which might lend some insight into this problem...

[16:16] Vincent Nacon: you need to provide a proper physic anyway

[16:16] Davido Chrome: Yes please!

[16:17] Vincent Nacon: .....you can do that with mesh

[16:17] Vincent Nacon: or a cube prim

[16:17] Sebastean Steamweaver: If you're not using mesh, it's usually a bad idea to set to physics_shape_none, because it forces the prims to be accounted according to the mesh standards, which really pumps up the Li.

[16:17] Andrew Linden: ground vehicles keep track of several "average collision normasl" where they hit the ground

[16:17] Andrew Linden: and build up a history of what that collision normal is at each point where the vehicle is colliding.

[16:18] Vincent Nacon: something like this shape would help you

[16:18] Rex Cronon: how long is the history?

[16:18] Andrew Linden: If they hit a momentary normal that is totally out of agreement with their history they will discard it, but they start to modify their history to accept such disagreeing normals.

[16:19] Andrew Linden: The idea being to discard the momentary incorrect normals encountered at prim seams at roads.

[16:19] Vincent Nacon: or edges

[16:19] Andrew Linden: I think the more collisions you get the more data goes into recomputing the expected normals

[16:20] Maggie Darwin: Two crashes in twenty minutes, then "you're trying to log in from another location so we're closing your viewer" to add insult to injury.

[16:20] Davido Chrome: Ah, so by optimizing I giving it less Data to use...

[16:20] Vincent Nacon: Maggie... just uninstall SL's viewer and install Firestorm

[16:20] Vincent Nacon: :)

[16:20] Rex Cronon: this can lead to problems. if u drive a car and bump into somthing, it can discard it

[16:20] Maggie Darwin: Not happening.

[16:20] Andrew Linden: so it makes sense that a "less cheap" vehicle that actually has more collision points will correct faster when encountering true ramps and obstacles (as opposed to bad normals at prim seams)

[16:21] Andrew Linden: spheres create the fewest collision points

[16:21] Andrew Linden: so I guess the first thing to try is to replace the spheres at the wheels with cylinders

[16:21] Vincent Nacon: that's why this prim is suggesting a slope in front of the wheel to expect coming ramp

[16:21] Vincent Nacon: no Andrew

[16:21] Andrew Linden: which should generate several more collision points

[16:21] Vincent Nacon: not cylinder

[16:21] Sebastean Steamweaver: My goodness! I'm not the only horse around anymore!

[16:22] Andrew Linden: or you could try Vincent's suggestions. I think he's well versed at tweaking vehicles.

[16:22] Fractal v2.98 whispers: This bike is locked.

[16:22] Vincent Nacon: cylinder is about the same as sphere if you're fixed on a one axis side

[16:22] Davido Chrome: I think I might. Thanks Vincent.

[16:22] Vincent Nacon: np

[16:23] Vincent Nacon: cylinder may be easier in collision from the sides but not in front of it

[16:23] Davido Chrome: If you make the cylinder or Sphere oblong it should stop being a calculated as a primitive though?

[16:23] Andrew Linden: I'm coming at it from a theoretical direction. I'm sure Vincent's knowledge is more experimental, so he probably knows what works well.

[16:24] Andrew Linden: Yes Davido

[16:24] Andrew Linden: er... it is still a "primitive" but does not collide as an "implicit" sphere

[16:24] Vincent Nacon: let's just say... long as you keep the shape more simpler than cylinder or sphere.

[16:24] Vincent Nacon: the better result

[16:24] Andrew Linden: it collides as a convex set of triangles

[16:25] Kandi Vendetta: What in moses ghost?

[16:25] Vincent Nacon: the top half of the bike doesn't really need to be simple as a box, it can be anything you wishes.... but the lower half is inportance

[16:25] Equino Faulkland: this vehicle conversation just got some more... horse power. >.>

[16:25] Vincent Nacon: importance*

[16:25] Sebastean Steamweaver chuckles.

[16:25] Davido Chrome: Sounds like what you really want is a golden middle way. Not too cheap, but cheap enough to cause less lag.

[16:25] Vincent Nacon: like this is good enough

[16:26] Ardy Lay: Put it in a john-boat

[16:26] Ardy Lay: All the works is a fluid simulation. ;-)

[16:26] Ardy Lay: -works +world

[16:26] Davido Chrome: Would probably work, yes.

[16:27] Ardy Lay: I had some fun trying to make a slide that depended entirely on physics to work.

[16:27] Vincent Nacon: the most common collision collider sides.... the simpler shape it should be. So simple like a rough cube.

[16:27] Ardy Lay: One of the service changes broke it. I should rebuild.

[16:27] Rex Cronon: cube wheels?

[16:28] Vincent Nacon: long as you have a slope in front of it

[16:28] Simon Linden: They don't actually roll in anycase

[16:28] Vincent Nacon: yup

[16:28] Davido Chrome: It does simcrossing much better than other bikes though, hacen't figure that one out. It's still 32 prims...

[16:28] Johan Laurasia: right, and that would be cool... jointed prims

[16:28] Vincent Nacon: they don't roll, just visual banking

[16:28] Simon Linden: How's the script count on it?

[16:29] Davido Chrome: The banking is real. These scripts use Angular motor for banking.

[16:29] Vincent Nacon: ....

[16:29] Davido Chrome: 2 scripts.

[16:29] Vincent Nacon: I wouldn't suggest doing a real banking for bikes...

[16:29] Simon Linden: That'll definitely help your region crossings

[16:29] Slice of Pizza ! whispers: MMMmmmm !

[16:30] Andrew Linden: right, fewer scripts should translate to better region crossings

[16:30] Vincent Nacon: but yeah... we need new vehicle API meeting some day

[16:30] Davido Chrome: Ah, yes. The current KCP scripts are like 6. I never did get why... I just took her basic scripts and modded them unrecognizable.

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

[16:30] Vincent Nacon: when Falcon isn't locked up in his cage

[16:30] Vincent Nacon: muhaha!

[16:31] Ardy Lay: How is the inventory work coming?

[16:32] Simon Linden: I'm clueless about that status, Ardy

[16:33] Andrew Linden: What inventory work? I think there is a team working on inventory but I don't know the details of it.

[16:33] Sebastean Steamweaver: I was curious as to whether or not any further discussion had been done on implementing llMatchGroup, or a possible prim-face float text replacement.

[16:33] Jonathan Yap: Direct Delivery

[16:33] Ardy Lay: Oz sent out notice to the opensource-dev mailing list some time ago telling third-party developers they should integrate a change.

[16:33] Andrew Linden: What are your least favorite inventory bugs that you want fixed Ardy?

[16:34] Kadah Coba: Its the Direct Delivery stuff

[16:34] Jonathan Yap: Slow loading

[16:34] Vincent Nacon: oh you mean inventory to sell over Marketplace, Kadah?

[16:34] Kadah Coba: The other was a cap added for Copy and Wear to resolve a race condition, I think

[16:34] Ardy Lay: Andrew, the only inventory bug bothering me has a work around. It's the bug that allows a resident to cause items to lose parenting when worn from the ground.

[16:34] Kadah Coba: Yes, Vincent

[16:35] Vincent Nacon: the "anti-magic box"

[16:35] Davido Chrome: Lose parenting?

[16:35] Kadah Coba: The first part of the Direct Delivery is the "everything you are given goes in to an new inbox system folder"

[16:35] Andrew Linden: Yeah, I saw that thread go by about the patch that should be incorporated but I forget the details...

[16:36] Andrew Linden looks in email.

[16:36] Kadah Coba: Which is apparently having some issues with the whole Copy and Wear thing

[16:36] Ardy Lay: https://lists.secondlife.com/pipermail/opensource-dev/2012-January/008536.html

[16:36] Jonathan Yap: That Received Items folder in the LL viewer is a huge "you cannot miss it" button at the bottom of the inventory list

[16:37] Ardy Lay: https://jira.secondlife.com/browse/VWR-25936

[16:37] JIRA-helper: [#VWR-25936] Attaching in-world objects consistently causes objects to appear outside of My Inventory

[16:38] Sebastean Steamweaver has a folder that's been stuck in his inventory for ages that way.

[16:38] Jonathan Yap: Firestorom has a fix for that, right?

[16:39] Vincent Nacon raise his shoulders with his hands

[16:39] Ardy Lay: If you follow the whole " Inventory Patch you should integrate " thread you find that there is some concern expressed by at least one of the devs that there may be a serious problem.

[16:39] Ash Qin: Shame that didn't make a jira on it.

[16:40] Ash Qin: Which is what the instructions on the wiki showed.

[16:40] Kallista Destiny: then inventory out of inventory requires , to the best of my knowledge, an intrvention by LL.

[16:41] Jonathan Yap: That intervention would be a good thing to have on a web page, you enter your details and some automatic process takes over

[16:41] Kallista Destiny: that is things that are outside of inventory (neither under the inventory , nor the library folder)

[16:41] Andrew Linden: Ah you're talking about an email Heni Beauchamp sent out yesterday?

[16:41] Andrew Linden tries to read that email...

[16:41] Rex Cronon: maybe u can tell on such web page to move/copy/delete what items

[16:42] Rex Cronon: automatic processes r not the best

[16:42] Jonathan Yap: There is a jira to make some operations into self-sufficient web services, to not involved support

[16:42] Ardy Lay: Andrew, yes.

[16:42] Rex Cronon: is like forcing an elefant into a speedo:)

[16:43] Andrew Linden: I see. Our plan is to force new inv items to show up in a particular inv folder

[16:44] Andrew Linden: (which, as a server-dev, I can see the appeal)

[16:44] Ardy Lay: Heh

[16:44] Rex Cronon: why not let use to set that folder?

[16:45] Rex Cronon: let user*

[16:45] Ardy Lay: Oh, the immutable system folder issue I was having I fixed myself by changing the viewer a little.

[16:45] Simon Linden: Yeah, it should get rid of the delay when buying stuff on the marketplace, plus vendors can deliver unpackaged items so there will be less wearing of boxes on your head

[16:45] Vincent Nacon: but I love wearing boxes on my head

[16:45] Rex Cronon: :)

[16:45] Johan Laurasia: I thought wearing a box was some sort of right of passage in SL

[16:45] Jonathan Yap: Is all new inventory going to go into the REceived Items folder? (that would make the most sense)

[16:45] Davido Chrome: Thank you soo much for the help. I will update my forum post with this new info once the transcript comes out. Now I'm gonna visit a friend in SL and then aknowledge the fact that its 2am in my part of the world and go to sleep...

[16:46] Keli Kyrie: but newbies with boxes on thier heads is so SL

[16:46] Johan Laurasia: yeah

[16:46] Vincent Nacon: used to be, now it's just goes to your hand

[16:46] Vincent Nacon: hand is the new default

[16:46] Andrew Linden: Well, I'm not 100% conversant with the plans for this project, but I can see the need/desire to put all new items in a known place.

[16:46] Simon Linden: yeah, hopefully soon we'll be able to tell them how bad it was in the old days

[16:47] Johan Laurasia: but I understand, I was talking to a newbie that thought he got ripped off because he thought the box it came in was just a picture of what he bought, the "box" it came in was flat, and had the product image

[16:47] Johan Laurasia: on it.

[16:47] Simon Linden: Jonathan - yes, so you can find stuff from marketplace or what other people give you

[16:47] Kallista Destiny: lols

[16:47] Johan Laurasia: I set him straight and showed him how to unpack it, but yeah, it's a newbie issue.

[16:47] Equino Faulkland: I might be in the wrong setting, but is this an appropriate forum to bring up issues with concierge live help?

[16:47] Andrew Linden: It seems to me that it should be possible for things to arrive in your inventory before your viewer is aware of it... they have to go somwhere.

[16:47] Johan Laurasia: maybe cover that sort of thing on orientation island

[16:48] Simon Linden: Not really, Equino. We're server developers here and pretty far from that side of things

[16:48] Andrew Linden: Once your viewer hears about the new objects it would be possible to sort them into other locations (perhaps even automatically).

[16:48] Ash Qin: Equino, avoid the scouts.

[16:48] Johan Laurasia: including NUL Andrew...lol

[16:50] Simon Linden: I think the idea is to make you sort them, so you hopefully remember where they go :)

[16:50] Jonathan Yap: Simon, are you going to be making a server change to send all new inventory to Received Items, or did you just think that was a good idea?

[16:50] Equino Faulkland: scouts?

[16:50] Jonathan Yap: (Maybe I need to write a jira)

[16:50] Andrew Linden: I see, the problem Henri is reporting is that the new items are supposed to be worn as soon as you get them - via the "copy and wear" operation

[16:51] Simon Linden: I'm not sure what happens server-side for that, Jonathan, but probalby not much. I think the new folder is just a special one everyone has, and new stuff gets put there via the database

[16:51] Jonathan Yap: Simon, I was thinking in terms of inventory arriving from objects, not the web store

[16:52] Andrew Linden: where does an llGiveInventory() show up now?

[16:52] Kallista Destiny: The still need to go to the inventory servers

[16:52] Ardy Lay: Objects

[16:52] Andrew Linden: Objects if it is an object inventory item, and Notecards if it is a notecard?

[16:53] Ardy Lay: llGiveInventoryList() makes a folder, I think

[16:53] Simon Linden: hmm, yeah, I'm not sure about that, but it would probably be easy ... just set the folder for the item to be the new special one

[16:53] Ardy Lay: Andrew, yes

[16:53] Draconis Neurocam: yes it does ardy

[16:53] Jonathan Yap: If it is just 1 item I think it goes into the folder of the right type, i.e. landmarks, scripts, etc

[16:54] Ardy Lay: I tend to keep my Objects folder clear of objects. It's filled with folders that I sort stuff into.

[16:54] Johan Laurasia: I do too

[16:54] Kallista Destiny: My notecards folder is a disaster, with objects not much better

[16:54] Simon Linden: Mine's full of things called "Object"

[16:54] Sebastean Steamweaver: I attempted to do that at one point, but It's still quickly gotten a mess.

[16:54] Johan Laurasia: it still tends to collect objects that need to be sorted though

[16:54] Ardy Lay: The Recent filter helps track down stuff.

[16:54] Johan Laurasia: lol, simon

[16:54] Johan Laurasia: the cure is to change the default name...

[16:55] Kallista Destiny: Speaking of server side, SVC-7443 seems to have been stealth implemented.

[16:55] JIRA-helper: http://jira.secondlife.com/browse/SVC-7443

[#SVC-7443] Telehubs overridden if teleporting a friend/avatar to a region where telehub exists outside of the telehub area

[16:55] Sebastean Steamweaver: It's difficult when you're uploading and copying and taking in objects all the time.

[16:55] Johan Laurasia: to "Thing" or something like tha.

[16:55] Ardy Lay: Simon, go to a protected snadbox and dump it on the ground.

[16:55] Kallista Destiny: OY!

[16:55] Rex Cronon: if u could see the date that was put into inventory might help sorting

[16:55] Andrew Linden: Henri also makes the point that the "Received Items" folder just adds yet another special, unremovable folder from your main inventory folder.

[16:55] Johan Laurasia: YAIS

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

[16:55] Johan Laurasia: I 2nd that one Rex

[16:55] Simon Linden: Yes Kallista, that's fixed in the RC channels now

[16:55] Jonathan Yap: The LL viewer now doesn't show empty system folders

[16:56] Kallista Destiny: well I donlt recall any discussion in these meating about it.

[16:56] Johan Laurasia: Well, at the end of the day, regardless, aren't we just sloshin' round keys here?

[16:56] Kallista Destiny: meetings*

[16:56] Ash Qin: Wasn't it brought up a month ago?

[16:56] Simon Linden: I don't either ... we probably found it with our regular bug triage and put it on the work list

[16:56] Andrew Linden: However I think making inventories much more flexible (so that the various "default" folders could be anywhere) would be hard to do.

[16:56] Kuroyu Canis: There are far too many 'special' folders as Is. I would break my arm off and GIVE it you if you could make it so all the special folders are hidden in one main folder and all you're inventory in another. or if you would jsut make it so we can move them. Good lord please don't add more Lol

[16:57] Sebastean Steamweaver: I did have a quick "just for info" question, if we have time for it Andrew.

[16:57] Andrew Linden: Go ahead Sebastean

[16:57] Sebastean Steamweaver: No need for extensive discussion really - I'm mostly just curious. I was curious as to whether or not any further discussion had been done on implementing llMatchGroup, or a possible prim-face float text replacement?

[16:57] Andrew Linden: llMatchGroup()? That doesn't ring a bell.

[16:57] Kallista Destiny: The problem is that 'fixing' this breaks established behaivior, since 2008 sy time I think.

[16:58] Simon Linden: What would that do?

[16:58] Ash Qin: I don't know, system folders serve a purpose. If you aren't using them, they don't appear in the viewer, so, you're not being forced to look at them if you aren't using them.

[16:58] Kuroyu Canis: llMatchGroup <--- Need it.

[16:58] Andrew Linden: what would llMatchGroup() do?

[16:58] Kallista Destiny: it matches the group of the object agains ANY group the the touching avatar is a memebr of

[16:58] Sebastean Steamweaver: llMatchGroup would allow a script to check and see if an avatar was a member of a specific group (by group key). It could only be detected if the group was (A) Active on the avatar or (B) publicly visible (not hidden).

[16:58] Kallista Destiny: not just the active group.

[16:58] Draconis Neurocam: matchgroup takes an avatar key and an object and checks if they are set to the same group

[16:59] Simon Linden: ah, I can see that would be useful, and a pain to implement ... like when you pick a group with 12k members in it

[16:59] Sebastean Steamweaver: Unlike llSameGroup, llMatchGroup could use the group key, rather than the group it's set to.

[16:59] Andrew Linden: so... integer is_in_group = llMatchGroup(agent_id, object_id); ?

[16:59] Sebastean Steamweaver: And it could check against visible groups, and not just the one active on the agents.

[17:00] Draconis Neurocam: err sorry it takes a group key and an avatar key, not an object key

[17:00] Draconis Neurocam: sorry

[17:00] Simon Linden: err, maybe not - the region would have to check the AVs group list, I guess

[17:00] Andrew Linden: hrm...

[17:00] Johan Laurasia: that sounds like what they're describing

[17:00] Sebastean Steamweaver: integer is_in_group = llMatchGroup(agent_id, group_id); ?

[17:00] Kelly Linden: Pretty sure there is an SCR for it.

[17:00] Andrew Linden: would you be wanting to use that on agent's that are offline?

[17:00] Draconis Neurocam: yes

[17:00] Sebastean Steamweaver: I personally wouldn't

[17:00] Kallista Destiny: thi way doors could have a list og groups for which they would open.

[17:00] Andrew Linden: or only agents that are present?

[17:00] Kuroyu Canis: Yes

[17:00] Draconis Neurocam: SCR-79

[17:00] JIRA-helper: http://jira.secondlife.com/browse/SCR-79

[#SCR-79] llMatchGroup() - Checks if an object or agent is active in a specified group.

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

[17:00] Johan Laurasia: he's saying an integer would be returned that would be TRUE or FALSE

[17:00] Sebastean Steamweaver: I think it should be used like llSameGroup, where it's only on avatars within a sim.

[17:00] Johan Laurasia: give that avatar & group id

[17:01] Squirrel Wood: Right now an object can only check if something is in the SAME group that the object is set to. But not, if that something is a member of the same group but just has another group set active

[17:01] Johan Laurasia: given*

[17:01] Andrew Linden: doing that query for offline agents is pretty hard to do... the simulator won't necessarily have the answer when you ask for it

[17:01] Simon Linden: It would have to be a command that triggered an event -- if the AV or group is not loaded already on the region, that data has to be fetched

[17:01] Andrew Linden: so it could take seconds before the answer could be fetched

[17:01] Johan Laurasia: yeah, but the query would be triggered by a touch

[17:01] Draconis Neurocam: people would be happy if it only worked for agents in sim, anything that allows something to check against a group the prim is not set to has been long lacking

[17:01] Kelly Linden: o/ Have a good weekend all.

[17:01] Johan Laurasia: afaik, offline avies don't touch objects

[17:01] Johan Laurasia: lol

[17:01] Kallista Destiny: right now there i an ugly hack involving haveing a stack of linked objecst each belonging to a different group and havine alll of them check.

[17:02] Andrew Linden: Yeah, handling offline agents would require a new event callback

[17:02] Simon Linden: Then there are privacy issues too

[17:02] Johan Laurasia: in what way?

[17:02] Andrew Linden: Are some group memberships private?

[17:02] Kallista Destiny: active avatars, in in the region only I think.

[17:02] Johan Laurasia: ah

[17:02] Johan Laurasia: yeah

[17:02] Simon Linden: YOu can query to see if I'm in all sorts of interesting groups :)

[17:02] Kallista Destiny: Yes you can hode membrship in a group.

[17:02] Johan Laurasia: those could just not be returned TRUE

[17:02] Johan Laurasia: even if true

[17:02] Sebastean Steamweaver: Some groups on a profile are "hidden" which is why I specified, it can only detect that if the group isn't hidden, or if it's currently active.

[17:03] Rex Cronon: i might be quite embarassing for some people if others found what xxx groups they belong to:)

[17:03] Simon Linden: and I want to publicly show my membership in the group

[17:03] Draconis Neurocam: it would be fine if it only worked to the current group title an avatar is set to

[17:03] Simon Linden: I don't want my meeroo group memberships leaking out

[17:03] Kallista Destiny: Alos in RP it prevents peeking at a profile to se if someone is an emey or a friend.

[17:04] Andrew Linden: Well... handling llMatchGroup() for avatars that are in the region, or can see into the region, would be pretty easy to do (assuming the hidden bit is already in simulator RAM (I don't know if that is the case yet))

[17:04] Sebastean Steamweaver: I wouldn't say it prevents that. All someone has to do is open the profile if they intend to do that.

[17:04] Andrew Linden: but support for offline avatars would be hard

[17:04] Ardy Lay: Heh, I joined a group with tens of thousands of members and the membership list kills the viewer.

[17:04] Rex Cronon: profile doesn't show hidden groups

[17:04] Kallista Destiny: The question was hidning groups.

[17:04] Andrew Linden: I wonder if there would be a way to implement the easy one first and then provide the offline stuff at a later date.

[17:05] Simon Linden: I have to run ... thanks everyone for coming and the good discussion

[17:05] Sebastean Steamweaver: That would be perfectly fine with me.

[17:05] Draconis Neurocam: yes same

[17:05] Rex Cronon: tc simon

[17:05] Draconis Neurocam: take care simon

[17:05] Keli Kyrie waves to Simon

[17:05] Simon Linden: See you all next week

[17:05] Sahkolihaa Contepomi: See you Simon & Andrew. Enjoy your weekends.

[17:05] Kuroyu Canis: bye simon :>

[17:05] Rex Cronon: u list all the groups on your profile:)

[17:05] Jonathan Yap: If you are offline you cannot be touching an object to trigger the event

[17:05] Kallista Destiny: I donlt think that the isues was deigned for offline Avatars

[17:05] Sebastean Steamweaver: I think privacy is respected so long as the script can't check any groups except that which are active or visible.

[17:05] Johan Laurasia: Andrew, I think the function is intended to be triggered by some sensor or touch or collision, so why would offline or even off region considerations would there be?

[17:05] Kallista Destiny: Exactly what Johan said

[17:05] Rex Cronon: put all groups u belong on a web page

[17:05] Johan Laurasia: if someone specifically set an offine avatars key, it could just return nothing.

[17:06] Jonathan Yap: You could go offline while the event was being handled

[17:06] Kallista Destiny: then it fails

[17:06] Andrew Linden: I'll have to read the jira...

[17:06] Johan Laurasia: yeah, but at the moment the check is being done, you're either on or offline.

[17:06] Johan Laurasia: oh I C

[17:06] Ardy Lay: Servers keep the infor for a bit after you go offline or leave the region.

[17:06] Johan Laurasia: duh

[17:06] Johan Laurasia: yeah, you could check while they were online.

[17:06] Kallista Destiny: and if you;r offline you didn't touc it, modulo some serious lag

[17:07] Ardy Lay: GHOSTS, again

[17:07] Kallista Destiny: laughs

[17:07] Andrew Linden: Well, I'll try to read the jira later and provide an more informed opinion next week.

[17:07] Sebastean Steamweaver: Thanks Andrew :) I appreciate the discussion

[17:07] Draconis Neurocam: indeed

[17:07] Andrew Linden: As to Sebastean's second item... per-face floating text?

[17:07] Johan Laurasia: Dito

[17:07] Andrew Linden: What is that one about?

[17:07] Andrew Linden: I thought we had HTML on a face these days.

[17:07] Sebastean Steamweaver: Oh, text on a prim face that doesn't require setting up an http_server.

[17:08] Squirrel Wood: one use as far as I am concerned would be to check if someone is in one of those pyramid scheme vampire groups and kick them off the sim... lest they start harassing people by attempting to "bite" them ...

[17:08] Andrew Linden: Is Kelly still here?

[17:08] Andrew Linden looks around...

[17:08] Johan Laurasia: that's what I was thinking, why not text on a prim.

[17:08] Squirrel Wood: Kelly left

[17:08] Sebastean Steamweaver: No, it seems he disappeared

[17:08] Ardy Lay: http://wiki.secondlife.com/wiki/Data_URI

[17:08] Sebastean Steamweaver: Basically, I'm sure yuo're aware of xY text?

[17:09] Sebastean Steamweaver: Yes Ardy, I want to aviod "shared media" for this at all costs.

[17:09] Draconis Neurocam: you can do some mean xytext with mesh

[17:09] Andrew Linden: XY text is a feature of OpenSim?

[17:09] Johan Laurasia: xy text is a bit primmy, and text on a prim is so much better.

[17:09] Sebastean Steamweaver: People should not be required to enable media in order to read text on a prim, and it can't be updated quickly.

[17:09] Ardy Lay: That method doesn't need an HTTP server.

[17:09] Sebastean Steamweaver: xY text is a script, a very heavy script.

[17:09] Ardy Lay: Hm.

[17:09] Johan Laurasia: not to mention a pain to output to..

[17:09] Andrew Linden: oooh XyText (By Xylor Baysclef?)

[17:09] Jonathan Yap: XY text is a kind of tortured prim so many faces show 1 way, then you apply letter textures to them

[17:09] Sebastean Steamweaver: It uses overlayed prims and texture offsets to build text.

[17:09] Sebastean Steamweaver: Yes

[17:10] Sebastean Steamweaver: Right now, other than floattext, it is our only option.

[17:10] Andrew Linden: yes I'm familiar with that

[17:10] Sebastean Steamweaver: And shared media

[17:10] Rex Cronon: u can do that better with sculpties/meshes

[17:10] Kallista Destiny: good night all

[17:10] Squirrel Wood: xyzzytext.. old. outdated.

[17:10] Rex Cronon: tc all those leaving

[17:10] Ardy Lay: Heh... I saw one that swapped sculpt maps around in prims to make letters.

[17:10] Vincent Nacon: I'm surprised no one done it with mesh yet... but I did for my pixel board

[17:11] Sebastean Steamweaver: But, shared media, while nice, is "too heavy" for things that don't need an implementation so involved. It would be preferable if people didn't have to enable media so that regular signs could be made, and you wouldn't need to worry about whether or not someone saw it.

[17:11] Ardy Lay: That got scary fast.

[17:11] Squirrel Wood: NexiiText uses the same principle as xyzzytext, but optimizes prim usage

[17:11] Johan Laurasia: yeah, it was quite useful in it's day, but I'd rather use text on a prim

[17:11] Andrew Linden: the Data URI feature doesn't work because why?

[17:11] Vincent Nacon: and this is made with Mesh

[17:11] Squirrel Wood: for text on a prim to work you need to have certain client features enabled.

[17:11] Sebastean Steamweaver: It requires the user to enable media, and not everyone has that feature turned on. It also can't be updated quickly.

[17:11] Squirrel Wood: I have them turned off by default.

[17:12] Sebastean Steamweaver: For example, script readouts that use float text can update very quickly - that isn't possible with shared media

[17:12] Andrew Linden: Ok

[17:12] Johan Laurasia: I was just thinking about that mesh sighn

[17:12] Sebastean Steamweaver: I also text to have them off by default.

[17:12] Vincent Nacon: and the blue one is old prim version

[17:12] Sebastean Steamweaver: also have*

[17:12] Andrew Linden: Sebastean, do you have an API proposal for your "text on a face" idea?

[17:12] Sebastean Steamweaver: There was one on JIRA actually.

[17:13] Andrew Linden: That is, an outline of how it would work?

[17:13] Andrew Linden: Ok, that number would be helpful.

[17:13] Sebastean Steamweaver: I'm not sure if it's still htere after the JIRA update, I'll take a look

[17:13] Squirrel Wood: notecard to "text on face" ?

[17:13] Andrew Linden: It seems to me that anything that put text on a prim face would need lots of control options

[17:13] Sebastean Steamweaver: https://jira.secondlife.com/browse/SVC-4715

[17:13] JIRA-helper: [#SVC-4715] llRenderText - Rendering Text on a Prim

[17:13] Andrew Linden: so you could get the text to look just right

[17:14] Vincent Nacon: yeah, I was kinda suggesting pixels!

[17:14] Sebastean Steamweaver: It depends - that proposal has a lot of options.

[17:14] Rex Cronon: we could use SVG for text on prim faces:)

[17:14] Ardy Lay: llLaTeX()

[17:14] Ardy Lay: Hehe

[17:14] Andrew Linden: LoL

[17:14] Sebastean Steamweaver: But honestly?I would be happy with even a barebones version that just controlled color and alpha.

[17:14] Vincent Nacon: but yeah... upload font? I sense a lot of legal copyright

[17:14] Rex Cronon: svg=scalar vector graphics

[17:14] Sebastean Steamweaver: I'm not concerned about having different fonts or bold or italics.

[17:15] Andrew Linden: sure, but maybe scrolling text?

[17:15] Andrew Linden: and Blink!

[17:15] Ardy Lay: SVG works already, in Data URI

[17:15] Sebastean Steamweaver: As far as I'm concerned, it could be the same font as float-text, as long as it can be used on a prim face.

[17:15] Rex Cronon: data uri has problems

[17:15] Andrew Linden: Automatic font size handling? Wrapped text?

[17:15] Ardy Lay: Unfortunately the URI capacity is small

[17:15] Squirrel Wood: as far as I am concerned it should look the same on the prim for everyone - regardless of text settings

[17:15] Sebastean Steamweaver: Wrapped text and font size would probably be necessary.

[17:15] Vincent Nacon: yeah, maybe

[17:16] Jonathan Yap: what happens to the text when you resize the prim?

[17:16] Ardy Lay: glue smoke

[17:16] Ash Qin: The other problem wtih using media on prim is that it forces the object to render if it's in the person's draw distance. If you've got a sim making use of media on prim for signs, you're going to have essentially 30 webpages open in the background to display text.

[17:16] Vincent Nacon: scale along with the surface?

[17:16] Sebastean Steamweaver: The font size would stay the same and the the basic unit size would be used to adjust the wrap text length.

[17:16] Andrew Linden: thanks for that jira link Sebastean.

[17:16] Ash Qin: Which is kind of ridiculous when all you want to do is render some text on a face.

[17:16] Sebastean Steamweaver: Or, the font could scale

[17:16] Andrew Linden: I'll try to read it before next meeting

[17:16] Sebastean Steamweaver: Either would work really.

[17:17] Andrew Linden writes these numbers down...

[17:17] Sebastean Steamweaver: Thank you again Andrew, I appreciate it.

[17:17] Vincent Nacon: but anywho, this meeting is long over

[17:17] Jonathan Yap: Thanks for staying late Andrew

[17:17] Sebastean Steamweaver: Have a good rest of your day

[17:17] Ardy Lay: Yeah, the current media rendering system is way to heavy for common text items.

[17:17] Ash Qin: Thanks Andrew.

[17:17] Vincent Nacon: take care all

[17:17] Rex Cronon: u should be paid overtime andrew:)

[17:17] Johan Laurasia: thanks Andrew, Simon

[17:17] Ardy Lay: I tried that and decided that the number of them should be kept very small.

[17:18] Keli Kyrie: Thanks Andrew. :)

[17:18] Favorites Drink Tray: Enjoy your Strawberry Daiquiri,Fine Caliber, so sweet and fruity!

[17:18] Andrew Linden: Thanks for coming everyone.

[17:18] Favorites Drink Tray: You DO know there's no tea in this, don't you, Fine Caliber?

[17:18] Rex Cronon: tc andrew

[17:18] Johan Laurasia: thank you

[17:18] Rex Cronon: tc everybody

[17:18] Ardy Lay: What I was thinking was it could be used in the viewer to create local texture assets



Simulator_User_Group

Prev 2012.02.07 Next 2012.02.14