Simulator User Group/Transcripts/2011.08.09

From Second Life Wiki
< Simulator User Group
Revision as of 14:49, 9 August 2011 by Andrew Linden (talk | contribs) (Created page with "Simulator_User_Group {| | Prev 2011.08.05 | Next 2011.08.12 |} == List of Spe…")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search


Prev 2011.08.05 Next 2011.08.12

List of Speakers

Andrew Linden Arawn Spitteler Draconis Neurocam
Helena Lycia Kadah Coba Kaluura Boa
Kelly Linden Latif Khalifa Leonel Iceghost
Liisa Runo Morgaine Dinova Pauline Darkfury
Qie Niangao Rex Cronon Sahkolihaa Contepomi
Vincent Nacon


[12:01] Helena Lycia: Hiya Andrew

[12:01] Draconis Neurocam: hello andrew

[12:01] Latif Khalifa: Hi Andrew

[12:01] Sahkolihaa Contepomi: Hey Andrew.

[12:01] Andrew Linden: Hello everyone.

[12:02] Morgaine Dinova: Hi Andrew

[12:02] Kadah Coba: Hai

[12:03] Andrew Linden: Hey I noticed a bug I haven't seen before... I wondering if anyone knows the easiest workaround...

[12:03] Kadah Coba: Hm?

[12:03] Vincent Nacon: which is?

[12:03] Andrew Linden: whenever I login these days I've got the "My Appearance" floater window open.

[12:03] Pauline Darkfury: hi folks :)

[12:03] Andrew Linden: And I can't seem to actually close it... only minimize it to the upper left corner.

[12:03] Sahkolihaa Contepomi: ...interesting.

[12:03] Kadah Coba wouldn't know.

[12:03] Vincent Nacon: that's new to me

[12:04] Vincent Nacon: but then, I don't use Viewer 2.x code

[12:04] Draconis Neurocam: i would try clearing your settings file

[12:04] Andrew Linden: I'm guessing a complete re-install of the SL viewer would fix it, or blowing away my history settings.

[12:04] Liisa Runo: Kelly, highlighting that one not going to work :P

[12:04] Draconis Neurocam: and let it rebuild it

[12:04] Latif Khalifa: very few non lindens here are on viewer2 so they woulnd't know

[12:04] Pauline Darkfury: it'll be part of the evil dock normally? Maybe you undocked it and need to find the non-intuitive arrow that re-docks it?

[12:04] Kelly Linden: ;)

[12:04] Andrew Linden: Yeah, ok. Just curious if any of you had seen it and figured out a different workaround.

[12:04] Draconis Neurocam: i use v2, whenever something odd happens like that its usually something wrong with settings.xml

[12:04] Rex Cronon: greetings all

[12:05] Morgaine Dinova: Hi Rex

[12:05] Sahkolihaa Contepomi always uses Snowstorm.

[12:05] Rex Cronon: hi morgaine

[12:05] Kadah Coba: The side bar is evil and its trying to steal your meeroos.

[12:05] Andrew Linden: Today I finished working on mesh parcel accounting... it works in my sandbox.

[12:05] Andrew Linden: We'll see.

[12:06] Morgaine Dinova: Any new things in the accounting?

[12:06] Andrew Linden: So I'm not exactly sure what my focus will be next, but I might be able to get some general maintenance work done soon.

[12:06] Pauline Darkfury: Ooooh, does that solve the new risks with over-full parcels?

[12:06] Kadah Coba: Improve concurrency performance.

[12:06] Kadah Coba: :3

[12:07] Latif Khalifa: Andrew, after last weeks meetings you asked to see the throne example I was mentioning I wonder what you make of those accounting numbers you saw there

[12:07] Latif Khalifa: I sent 3 different versions

[12:07] Andrew Linden: Yes Pauline, well... most of the issues of overfull parcels.

[12:07] Helena Lycia: Yikes

[12:08] Andrew Linden: There are a few very rare holes by which objects might be able to overfill a parcel, but the objects should now be returned in the "correct" order.

[12:08] Kadah Coba: Good to hear

[12:08] Pauline Darkfury: Great, thanks :) Can we please try to get that into mesh before mesh goes grid-wide

[12:08] Latif Khalifa: yeah there will be a lot of accidental returns when mesh rolls

[12:08] Andrew Linden: I put correct in quotes because it should work as intended, but it is always debatable as to what is absolutely the best order.

[12:09] Pauline Darkfury: As long as it confidently mitigates the new risks, such that the sudden bloat won't cause random returns for innocent victims, I'm very happy

[12:09] Andrew Linden: Yes, random returns are fixed. It should return the objects that pushed the parcel over first.

[12:10] Andrew Linden: And in most cases it will sucessfully pre-check the parcel owner resource limits before changes happen, and block them from succeeding.

[12:10] Andrew Linden reads back...

[12:10] Vincent Nacon: is it also tested with temp prims?

[12:10] Andrew Linden: lemme look at that image...

[12:11] Latif Khalifa: I sent you objects after the meeting as you have asked

[12:11] Latif Khalifa: 3 versions

[12:11] Andrew Linden: Oh I see. That throne has a number of concave parts.

[12:11] Pauline Darkfury: Even better. The bit that was really scaring me was 1 nooby tenant doing the wrong thing (or even a non-noob getting a very unexpected bloat), and kicking objects belonging to 20+ other innocent tenants back to them

[12:11] Andrew Linden: That's why the cost is coming out a little higher when you link a mesh.

[12:11] Latif Khalifa: "little higher" lol

[12:11] Latif Khalifa: 22 -> 212

[12:11] Rex Cronon: why don't u make it so that people can view how many prims their objects are before rezzing them. while still in their user directory?

[12:11] Andrew Linden: The "true" cost of that throne is higher than the 22 prims you started with.

[12:12] Helena Lycia: And I've got 25 + 2 = 250

[12:12] Andrew Linden: However, when you use legacy prims you get the artificially incorrect low cost.

[12:12] Helena Lycia: 450

[12:12] Pauline Darkfury: Latif, it's simple, you just need to make your throne 32x32, so that it occupies the correct amount of land for its prim usage :p

[12:12] Andrew Linden: No that won't help. Scaling up will raise the cost, when mesh is included.

[12:12] Morgaine Dinova: Pretty much every object of interest has concave parts. The set of objects without concave parts is, well, pretty uninteresting.

[12:12] Latif Khalifa: Andrew, I change all prims to non physics excapt one box, and I still get higer numbers.

[12:13] Helena Lycia: It just seems wrong when a seemingly innocuous build can have such a vast PE

[12:13] Latif Khalifa: So you are telling me that having 22 colliding prims is cheaper for the physics engine than 1 colliding box?

[12:13] Andrew Linden: Yeah Latif, I'd guess that you're hitting the "Streaming Cost" rather than physics.

[12:13] Morgaine Dinova: What is the definition of "streaming cost"?

[12:13] Latif Khalifa: Yes, that's exactly my point. The current accounting add ficticious cost

[12:14] Latif Khalifa: there is no straeming cost

[12:14] Andrew Linden: Disabling the shapes will help when the physics cost is the highest contributor.

[12:14] Helena Lycia: Agreed

[12:14] Helena Lycia: (with Latif)

[12:14] Pauline Darkfury: but that really doesn't make sense either, if it is mostly traditional prims. The streaming cost of them should surely almost always be 1.0?

[12:14] Kadah Coba: "Streaming" is replaceable with "render" in nearly all usages of "streaming costs"

[12:14] Draconis Neurocam:

[12:14] Andrew Linden: Yeah well, we're calling it "streaming cost".

[12:14] Latif Khalifa: Pauline, the streaming cost is 10 times the prim count

[12:14] Andrew Linden: You have to ask the people at the Mesh User Group to get an explanation of streaming cost.

[12:14] Morgaine Dinova: You're call *what* "Streaming cost"?

[12:14] Latif Khalifa: and that cost doesn't exist

[12:15] Andrew Linden: Perhaps I should attend and take notes so I can fully grok it myself.

[12:15] Vincent Nacon: it's still misleading

[12:15] Pauline Darkfury: yeah, that's what I'm saying - if they are simple prims, PE should be no worse than 1x<prims> if there's no evil physics in the picture

[12:15] Andrew Linden: I suspect that the streaming cost formulae are going to need some work.

[12:15] Kadah Coba: It seems to vary depending on the hour of the day, it either the rendering cost or the cost for the ammout of data that needs to be sent to the client.

[12:15] Helena Lycia: What I don't understand is why is it so high compared to pure prim builds of greater complexity?

[12:15] Latif Khalifa: The idea of "streaming cost" for regular prims is all wrong

[12:15] Latif Khalifa: who came up with it?

[12:16] Qie Niangao: So... is there some compelling reason to discuss non-physics Mesh PE here?

[12:16] Morgaine Dinova: Andrew: I don't go to the mesh meetings, but from reports it seems they're completely irrational. This is why I'm trying to get a rational answer from you. :-) Your team tends to stick to tech facts.

[12:16] Vincent Nacon: should have been something like Render cost and Data cost

[12:16] Pauline Darkfury: I can see >1.0 from something horrible in physics terms, but it's absurd if the physics is more or less nulled (or reduced to simple box/sphere)

[12:16] Helena Lycia: I get why a mesh torus will be slightly more than a prim torus but the inflation on some builds just seems wrong

[12:16] Vincent Nacon: Steaming is too confusing for the lesser skills builders

[12:16] Pauline Darkfury: it's confusing as well for those with skills, as far as I can see

[12:16] Latif Khalifa: In effect LL is charging land owners as if mesh was generated from prims server side and then transmitted

[12:16] Latif Khalifa: which is not the case

[12:17] Helena Lycia: It's worse than that

[12:17] Andrew Linden: Yeah it is confusing. I'll try to study it and come back to see if I can explain it.

[12:17] Vincent Nacon: ok correction, not confusing, just mind blown

[12:17] Morgaine Dinova: "Streaming cost" needs a precise definition related to actual streaming. Otherwise it needs to be called "fudge factor".

[12:17] Helena Lycia: I had a build that was 27 prims, I replace 2 prims with a 2 PE mesh and the result is a 450 PE object

[12:17] Latif Khalifa: So you get a situation that you want to help simulator by reducing the number of colliding prims and end up 10 doubling the prim count

[12:18] Andrew Linden: Maybe we could call it "Mystery cost" and that could clear things up.

[12:18] Kadah Coba: I mentioned as much to the mesh team, they looked at me as if I was crazy

[12:18] Vincent Nacon: I understand what LL is trying to do with prim counting... might be best not to call it PE or Prim for that matter anymore

[12:18] Kadah Coba: They said the costs were easy to understand

[12:18] Latif Khalifa: yeah the mesh team are useless

[12:18] Helena Lycia: In my case I was trying to remove two cut toruses from a physical object and replace it with a simplier shape that should have been kinder to physics but get something that can't even be physical

[12:18] Andrew Linden: I agree. I don't like the term "Prim Equivalent Cost"

[12:18] Morgaine Dinova: Andrew: Although "mystery cost" wouldn't clear anything up, it would avoid accusations of misrepresentation, because techs know what the actual costs of streaming are.

[12:18] Andrew Linden: I suggested earlier that we use "Units"

[12:19] Pauline Darkfury: There's one easy fix, "integer max_prims=150000;", problem solved, and we'll take our random prim counts happily ;)

[12:19] Vincent Nacon: something goes along with "resource"

[12:19] Andrew Linden: however it was not supported.

[12:19] Latif Khalifa: Can anybody at Linden Lab look into the justifiabilty of "streaming cost" for regular prims. It's going to cause a lot of damage.

[12:19] Kadah Coba: "Prim Exponential Cost"

[12:19] Andrew Linden: I think we need a name for the unit... maybe named after someone.

[12:19] Helena Lycia: Well "Prim Equivalence" is right if LL are going to use that in calculations of when a parcle or sim is "full"

[12:19] Vincent Nacon: just won't be easy to change their mind set about prim in general, best to start something new to make it easier gasps the new concept of resource management

[12:20] Draconis Neurocam: the M number >.>

[12:20] Rex Cronon: data weight?

[12:20] Vincent Nacon: yeah, something like that

[12:20] Morgaine Dinova: Name the units afte Zero Linden. Since N * Zero = 0, everyone should be happy :P

[12:20] Pauline Darkfury: that's the thing, it doesn't actually seem to be the data weight

[12:20] Vincent Nacon: or give it a new name

[12:20] Kadah Coba: I suggested that the server load costs should have been a secondary accounting number on parcels and sims. It might have been easier to understand

[12:20] Vincent Nacon: made-up name that sounds like a unit thing

[12:20] Latif Khalifa: "resource maagement" is fine. but charging $8 USD a month for that throne cannot be justified in any shape or form.

[12:20] Pauline Darkfury: I can get on board it having some sort of relationship with the actual size of data, plus the physics CPU guesstimate

[12:21] Andrew Linden: Personally I don't like the term "weight", because when talking about "physics weight" things start to become ambivalent.

[12:21] Andrew Linden: er... ambiguous I mean.

[12:21] Pauline Darkfury: This seems quite different to the actual data size/weight

[12:21] Kelly Linden: Then don't change the throne to the new accounting. Is there a reason you can't leave the throne as is?

[12:21] Latif Khalifa: there is no streaming cost to prims, construction data is what, 100 bytes?

[12:21] Morgaine Dinova: Although we're joking here, a nice unit name won't solve the problem. The weightings can't be onerous or you're killing SL.

[12:22] Rex Cronon: i missed lots of interesting stuff, because my connection broke:(

[12:22] Latif Khalifa: Kelly, I will not. But I would like to be able to make content that is less costly to the engine, by say using fewer colliding prims

[12:22] Pauline Darkfury: Yeah, prim data really should be tiny as far as the real streaming cost goes

[12:22] Andrew Linden: Well, Latif can leave the throne as is to use the legacy formulae, but it highlights the problem that others will discover for other content.

[12:22] Helena Lycia: Yeah, unless there are 2 quotas instead of one. A prim cost and another cost and the limit on the new cost has to be a lot higher

[12:22] Kelly Linden: If you make a visually equivalent chair, fully in mesh, it would be very useful to see what that cost ends up being.

[12:22] Vincent Nacon: it shouldn't be a prim since it's a new concept of data management

[12:23] Vincent Nacon: thus it's not a prim anymore

[12:23] Latif Khalifa: But imagine people linkin 1 mesh prim to their 50 prim house resulting in 500-1000 prim housr

[12:23] Pauline Darkfury: Yeah, the trouble with "just leave it on traditional accounting" is that it's worse for the physics load on the sim. Give us anything which comes out as less than 100% when we save some physics load, and we'll make SL perform better

[12:23] Helena Lycia: I really think mesh is going to end up being used almost exclusively for avatars because so much of what could have been is just going to be too costly

[12:23] Andrew Linden: Other news before the hour is up...

[12:24] Leonel Iceghost: the thing is that the throne is cheap properties, and a box for the physics right? yet it is more PE

[12:24] Andrew Linden: I didnt' have time to look into the "sim time stats not adding up" problem.

[12:24] Kaluura Boa: Prim Equivalence should just never apply to prims. That's stupid. 1 prim == 1 prim != x PE

[12:24] Morgaine Dinova: Any real streaming cost can't possibly be static anyway, since only viewers with zero-size cache give the sim a significant streaming cost. Actual asset streaming costs almost vanish with a large cache, which will come eventually.

[12:24] Andrew Linden: But I did talk to someone about enabling the prim encroachment returns on the mainland.

[12:24] Arawn Spitteler: Perhaps we should have honest accounting for Phys Cost, in stead of just Prim Counts.

[12:24] Pauline Darkfury: Well, I'd argue that 1 prim should be strictly <= 1.0 PE

[12:24] Morgaine Dinova: Pauline++

[12:24] Rex Cronon: u can say that 1prim = x nr of vertices

[12:24] Latif Khalifa: Pauline, yes

[12:25] Latif Khalifa: certainly not 10 or 20

[12:25] Andrew Linden: They identified a problem that needs to be fixed, but it shouldn't take long to get done.

[12:25] Draconis Neurocam: what was the problem out of curiousity?

[12:25] Andrew Linden: Then we have to wait for mesh to get everywhere, but I might be able to fix the problem shortly after it promotes.

[12:25] Andrew Linden: There is some special behavior for mainland regions for encroachment

[12:26] Vincent Nacon: I'm assuming you meant Havok 2k10 to go with mesh as well?

[12:26] Andrew Linden: Hrm....

[12:26] Pauline Darkfury: cool, any progress on mainland encroachment is welcome

[12:26] Qie Niangao: agreed

[12:26] Kadah Coba: Ditto

[12:26] Vincent Nacon: if so... density need to be fixed before then

[12:26] Andrew Linden: you know, I just relized that the problem may not be a problem after all!

[12:26] Pauline Darkfury: One thin to note on that subject, Jack's account has been renamed to Robin, and he still owns various lumps of LDPW type content

[12:26] Liisa Runo: what special behavior?

[12:27] Andrew Linden: I had forgotten one little detail... the special behavior I was talking about actually kicks in for any region for whom the estate owner is Governor Linden.

[12:27] Andrew Linden: Which would cover the problem estate (Homesteads) as well as mainland.

[12:27] Pauline Darkfury: It's probably all still in group Maint tho, so that could be the "easy" whitelist criteria

[12:27] Draconis Neurocam: parent estate 0 if i remember?

[12:28] Andrew Linden: Anyway, the special behavior is that some content owners are considered "Estate Mangers" for content ownership, and can be made immune to encroachment return.

[12:28] Andrew Linden: Some such owners are the "Maintenance Group" and "DFPW" (Dept. for Public Works)

[12:28] Pauline Darkfury: yup, please do make Maint group, Michael, Jack, the Moles, etc exempt

[12:28] Andrew Linden: No, not Moles, but Maint Group yes.

[12:29] Liisa Runo: what about linden trees and plants? some of them have lot bigger size in numbers than the actual visual size

[12:29] Pauline Darkfury: Well, I'd argue that Mole encroachment should be such an infrequent problem case, that it would be safer to just make them exempt and need a ticket (Micahel is speedy at LDPW tickets)

[12:30] Andrew Linden: Anway, I'm going to try to convince the right people to open up encroachment return for the mainland when mesh goes out.

[12:30] Meeter: Timecheck : User Group is half over

[12:30] Andrew Linden: I'm pretty sure encroachment return won't mean the end of the world as we know it, btw.

[12:30] Draconis Neurocam: i think it will be a much improved world

[12:30] Pauline Darkfury: No, I think it will be a huge improvement to quality of life on mainland

[12:30] Latif Khalifa: yep

[12:30] Vincent Nacon: wasn't worried about that, more about the density though

[12:31] Latif Khalifa: it's a welcome change

[12:31] Andrew Linden: Hrm... trees and plants.

[12:31] Pauline Darkfury: I can get mainland encroachment solved when it matters, but not via just an AR, it takes a good bit more effort than that to get it solved quickly

[12:31] Rex Cronon: i don't think that everybody would like it if the linden roads is encroaching on their land and takes away their prims

[12:31] Andrew Linden: I don't think we properly detect encroachment for trees and grass.

[12:31] Andrew Linden: I guess that could be considered a bug.

[12:31] Pauline Darkfury: encroachment doesn't take prims

[12:31] Vincent Nacon: maybe you'll get less ARs about plants

[12:32] Andrew Linden: We'll have to work on that when we know how bad it is.

[12:32] Qie Niangao: the dimensions of Linden plants are pretty funky numbers, I realize. Until I tried (unsuccessfully) to make 64m Linden trees, I never appreciated that Linden grasses start out about 30x30m.

[12:32] Vincent Nacon: oh

[12:32] Pauline Darkfury: Michael has a bit of road on my land, no prim count, and he's welcome to keep it there

[12:32] Pauline Darkfury: Equally, I have stuff connecting to road on LDPW land, but with the root prim on my land

[12:32] Vincent Nacon: actually, isn't encroachment based on Boundary Box?

[12:32] Andrew Linden: Right, well I think Michael's road will actually be immune to encroachment return.

[12:32] Latif Khalifa: yeah foliage is really a viewer side effect, i imagine it's tought to determine if it enroaches

[12:32] Pauline Darkfury: the road in question was there when I bought the parcel, so I knew it was always going to be there

[12:33] Andrew Linden: I believe Micheal Linden is included in the special list of "Linden Content" owners.

[12:33] Vincent Nacon: plants do have BB

[12:33] Vincent Nacon: should able to work

[12:33] Andrew Linden: Yes, plants have a rendering bounding box.

[12:33] Liisa Runo: they have BB, but it dont mathch the actual visual size

[12:33] Latif Khalifa: do they have server side scale?

[12:34] Latif Khalifa: and could scale be used as BB

[12:34] Andrew Linden: But did you know... plants also have a body in the physics engine! However, plant shapes are collisionless.

[12:34] Vincent Nacon: the BB is always bigger than the visual, shouldn't be a problem to return them

[12:34] Draconis Neurocam: interesting

[12:34] Andrew Linden: I added collision bodies to trees back in the Havok4 project so I could remove a bunch of special case code that would walk around the issue.

[12:34] Rex Cronon: isn't it useless their physics body?

[12:34] Pauline Darkfury: depends if the BB used by encroachment return is done the same way as the viewer's debug BBs

[12:34] Liisa Runo: problem become when annoying neighbor return tree that is barely even near the border

[12:34] Andrew Linden: Yes, useless. I had plans for making them optionally collidable.

[12:35] Andrew Linden: I think a few people would like to have their tree trunks collidable.

[12:35] Andrew Linden: However that work was never completed.

[12:35] Vincent Nacon: at last, Linden's Plant got some edges over prim version plants

[12:35] Vincent Nacon: muhaha!

[12:35] Pauline Darkfury: I'm generally very happy with the trees as-is, all phantom, tbh

[12:35] Andrew Linden: Also, I wanted to allow trees to link to other objects (potted tree, for example)

[12:35] Latif Khalifa: Liisa there are fascist neighbours like that in RL too :P

[12:35] Andrew Linden: and maybe even get scripted.

[12:35] Vincent Nacon: O_o to do what?

[12:35] Draconis Neurocam: ooh scripting them could be interesting

[12:35] Vincent Nacon: walk around?

[12:36] Kaluura Boa: Long live the walking trees!

[12:36] Andrew Linden: It never got done, but it is easier to do when they have a collison body in the phys engine.

[12:36] Pauline Darkfury: If you get a neighbour like that, Liisa, just remember that it works both ways, so do check all of their prims are not 1mm over too ;)

[12:36] Draconis Neurocam: you could make them turn toward the sun to simulate phototropism

[12:36] Liisa Runo: but if we kill a neighbor in RL, they actually die forever and ever. SL not that practical

[12:36] Helena Lycia: Yay - Linden Triffids

[12:36] Sahkolihaa Contepomi: Talking trees.

[12:36] Latif Khalifa: yeah, trees being phantom is fine i don't think you really need to bother with that

[12:36] Vincent Nacon: but anywho, have you talked to Falcon about prim's density issue?

[12:37] Vincent Nacon: he's really confused

[12:37] Morgaine Dinova: It's just like RL. Most politicians are vegetables too.

[12:37] Andrew Linden: What is the "prim density" issue? I think I saw that mentioned a little while ago.

[12:37] Andrew Linden: But I didn't understand it.

[12:37] Vincent Nacon:

[12:37] JIRA-helper: [#CTS-573] Incorrect Prim Density Value

[12:37] Rex Cronon: i am not sure about something. if u rez somethin on your parcel but it encroaches on your neighour parcel, wil the sim return it or that owners parcel will have to?

[12:37] Draconis Neurocam: im pretty sure he means the density with the new physics settings for 2k10

[12:37] Andrew Linden: Oh right, mass density.

[12:37] Latif Khalifa: Rex, the other parcel owner

[12:37] Andrew Linden: No I haven't spoken to Falcon about that yet.

[12:38] Latif Khalifa: nothing automatic (thanks god) :P

[12:38] Vincent Nacon: since the last revision, he only changed the label for the input

[12:38] Vincent Nacon: and the value is even worst

[12:38] Vincent Nacon: he made them 100,000x more

[12:38] Andrew Linden: Latif is correct, encroachment return is only a manual process. The sim won't do it automatically.

[12:38] Rex Cronon: wouldn't be easier if the sim didn't allow u to rez objects that encroach on others land?

[12:38] Latif Khalifa: good :)

[12:39] Andrew Linden: Rex, I think maybe we should add that as a per-parcel option.

[12:39] Andrew Linden: However, we'll see how it goes with manual intervention first.

[12:39] Vincent Nacon: the correct value for default should be 10

[12:39] Rex Cronon: ok

[12:39] Morgaine Dinova: {Andrew,Rex}++

[12:39] Qie Niangao: It might be nice for the object owner to know that part of the object is encroaching, so they know they're living in sin.

[12:39] Liisa Runo: i think manual is enough, automatic system will cause problems

[12:39] Latif Khalifa: Rex, there are many situations where enrachment is not a problem. Like when you are building a skybox it's hard to tell if you move the prim 1m over

[12:39] Pauline Darkfury: ugh, I'm not convinced by blocking encroaching rez. That will make stuff very annoying for some legit cases

[12:39] Latif Khalifa: manual is just fine

[12:40] Andrew Linden: Yes Qie, encroachment feedback would be nice.

[12:40] Vincent Nacon: not 1,000 by value of 100sq^3m (which is 100,000)

[12:40] Rex Cronon: right. at least a warning

[12:40] Pauline Darkfury: I.e. with rez box houses, you never know just quite where they are going to rez

[12:40] Pauline Darkfury: To me, it's legit to encroach for 5 mins here than there while positioning something like that

[12:40] Latif Khalifa: also for skyboxes people don't care if you enroach

[12:40] Pauline Darkfury: Some do, sadly, Latif

[12:40] Latif Khalifa: well most people

[12:40] Draconis Neurocam: i do, haha

[12:41] Latif Khalifa: (most sane people) :P

[12:41] Vincent Nacon: and please do something about it because I'm getting tired of bringing it up each time I attended

[12:41] Pauline Darkfury: Me, as long as the skybox isn't causing real issue, it can ecroach all it wants

[12:41] Morgaine Dinova: Latif: but if a sim has encroachment control, then you could simply push or expand your object as far as possible towards the parcel boundary, and the sim would stop you from going too far automatically.

[12:41] Pauline Darkfury: I happily let someone who has a 128 surrounded by me on all sides use a 32 or 40m scenic sphere on their micro skybox

[12:42] Latif Khalifa: I think manual control would do fine. I don't mind tiny bits protruding if it's not really bothersome.

[12:42] Pauline Darkfury: yeah, just being able to return the problem stuff without a major fight will be wonderful

[12:42] Pauline Darkfury: that's all we really need

[12:42] Latif Khalifa: and I was always lucky to have reasonable people around me. manual control for those who aren't is perfect :)

[12:43] Morgaine Dinova: Returning stuff creates fights. Encroachment control not letting you encroach in the first place would avoid fights. And avoid ARs, reduce support costs, and avoid bad feeling in general.

[12:43] Rex Cronon: i think morgaine is right

[12:43] Pauline Darkfury: No, I'd say that the rez-blocking would cause just as many fights

[12:44] Vincent Nacon: buy more land, get over it

[12:44] Draconis Neurocam: better scripted control over terrain would be nice, just a random idea

[12:44] Pauline Darkfury: it would create exactly the same bad blood that exists over ban lines at present, but bigger

[12:44] Andrew Linden: Yes. I think encroachment feedback (visible indication of encroachment as you place objects) would the next step.

[12:44] Rex Cronon: who r u going to fight with if the sim doesn't let u rez:)

[12:44] Qie Niangao: In practice, the mere *prospect* of things getting returned manually has already improved behavior that I've seen.

[12:44] Pauline Darkfury: That is frankly a disgusting answer, Vincent :(

[12:44] Morgaine Dinova: Vincent: spoken like a true rich person.

[12:44] Vincent Nacon: logical, not rich

[12:45] Liisa Runo: automatic system would effectively kill traveling with vehicles for example

[12:45] Morgaine Dinova: illogical and anti-social to boot

[12:45] Pauline Darkfury: If someone buys a 4096 and then buys a 60x60 house, it is not easy to position it. If they were surrounded by obnoxious control freaks, it would become almost impossible

[12:45] Vincent Nacon: what was the other option?

[12:45] Andrew Linden: The auto system could be connected to the autoreturn system.

[12:45] Andrew Linden: So encroaching objects could be returned on a timer.

[12:45] Andrew Linden: er... I mean the encroachment return system could be connected to autoreturn.

[12:45] Vincent Nacon: that could work

[12:46] Draconis Neurocam: the visual system could make use of parcel boundries, just make them more three dimensional rather than two

[12:46] Liisa Runo: you better add settings for my 3 years old client to allow me to let my neighbors tree poke to my parcel, i dont want some stupid automatic system to ruin my and my neighbors carefully planted border trees

[12:46] Pauline Darkfury: yeah, autoreturn type semantics would be tolerable, but I'd want it to be an optional feature. I insist on keeping autoreturn on to keep junk away, and wouldn't want that to kick out someone's scenic that was in no way bothering me, but was encroaching

[12:47] Qie Niangao: Pauline: right

[12:47] Andrew Linden: Yeah I know. It could only be added as an option.

[12:47] Rex Cronon: it should apply to things thing that u rez, not to what is already rezzed

[12:49] Liisa Runo: kinky people got forzed to move to zindra, i suggest that we make entirely new continent for land assholes like ban line people and automatic encoachment return people.

[12:49] Latif Khalifa: Liisa++

[12:49] Pauline Darkfury: YAY, let's have a Bandra continent! :)

[12:49] Andrew Linden: Hehe

[12:49] Morgaine Dinova: lol

[12:49] Draconis Neurocam: will happily move to the land of ocd people

[12:50] Andrew Linden: brb... 2 min

[12:51] Andrew Linden: back

[12:51] Vincent Nacon: wb

[12:51] Rex Cronon: wb

[12:51] Draconis Neurocam: andrew has mastered timewarp

[12:52] Helena Lycia: PE costs warp space-time

[12:52] Vincent Nacon: under a minute or badly guessed his time

[12:52] Andrew Linden: Well, that's all the news I've got for today. We need a new topic.

[12:52] Latif Khalifa: Helena lol

[12:52] Liisa Runo: new topic: meeroo's birth control?

[12:52] Latif Khalifa: they bend the bounds of reason too

[12:52] Pauline Darkfury: From timewarp convo in SL Beta earlier, a bunch of SL educators timewarping!

TimeWarp @ VWBPE Conference 2010

[12:52] Vincent Nacon whispers: so, will you see what you could do about the density then

[12:52] Vincent Nacon: ?

[12:53] Vincent Nacon: ahh silly whisper key

[12:53] Helena Lycia: Advice to Lindens - when PE costs are so high that they warp space-time , they're too high

[12:53] Andrew Linden: Vincent. I'll talk to Falcon about it.

[12:53] Pauline Darkfury: Or, The Sims timewarp:

Let's Do The Time Warp Again

[12:53] Andrew Linden: I think I know how the conversation is going to go...

[12:53] Vincent Nacon: and remember, prim's default density should be 10

[12:53] Andrew Linden: Actually, Vincent, do you have a proposal for how it should be fixed?

[12:53] Vincent Nacon: how? it's a input value

[12:54] Vincent Nacon: can't be more simple than that

[12:54] Draconis Neurocam: i think andrew means what the correct conversion should be, defaults etc

[12:54] Vincent Nacon:

[12:54] JIRA-helper: [#CTS-573] Incorrect Prim Density Value

[12:54] Vincent Nacon: = m / V where = density (kg/m3) m = mass (kg) V = volume (m3)

[12:55] Rex Cronon: u r getting physical vincent:)

[12:55] Vincent Nacon: I have to

[12:55] Vincent Nacon: :)

[12:55] Andrew Linden: The real problem is that the density units as used in the new UI are off by some factor from the units returned by llGetMass() in LSL

[12:55] Meeter: Timecheck : User Group is almost over

[12:55] Andrew Linden: so... we could change the units returned by LSL, or change the units shown in the UI

[12:55] Vincent Nacon: right but it's still 10 if you use LSL anyhow

[12:56] Vincent Nacon: yeah I think it's the unit in UI

[12:56] Vincent Nacon: nothing wrong with llGetMass or any other function/codes

[12:56] Andrew Linden: so, you would like to see the UI use the same units as LSL?

[12:56] Pauline Darkfury: Yeah, they do need to match, for the principle of least astonishment

[12:56] Vincent Nacon: aye

[12:57] Andrew Linden: oh and btw, the units expressed in LSL are way too light for "real life" mass/density units.

[12:57] Pauline Darkfury: If there's a reason that the number has to be odd in the protocol between viewer & server, maybe just scale them to be correct at display time

[12:57] Vincent Nacon: even though SL never had a proper unit, it still come up with a "10" of something

[12:57] Vincent Nacon: but the value in input is set at 1,000 by 100"unit"^3m

[12:57] Andrew Linden: ok, that's what I wanted to know

[12:57] Vincent Nacon: which is generally speaking it's 100,000

[12:58] Vincent Nacon: and yeah, we're lighter than gas

[12:58] Andrew Linden: maybe we just need to specify that the units returned by LSL are in 100k... except for the fact that the avatars are made artificially heavy, as I recall.

[12:58] Vincent Nacon: but the problem is some builders already use the density for their vehicles and some other content

[12:58] Vincent Nacon: may break some vehicles and such

[12:59] Andrew Linden: Is there a User Experience User Group in the schedule?

[12:59] Vincent Nacon: even if it's not gram or lindogram, whatever the unit is

[12:59] Pauline Darkfury: well, would it be so terrible to just document the differences on the LSL page, if the current viewer numbers are more "correct" in real terms?

[12:59] Draconis Neurocam: sadly no andrew

[12:59] Kaluura Boa: btw, any plan for the LSL function for density, friction and restitution? And slice?

[12:59] Andrew Linden: That is, a user group focused on User Interface?

[12:59] Draconis Neurocam: bambers said something about one

[12:59] Vincent Nacon: there was UI group

[13:00] Draconis Neurocam: and then i never heard of it ever again

[13:00] Vincent Nacon: not sure if that's still around since that one linden left the lab

[13:00] Meeter: Thank you for coming to the Server User Group

[13:00] Helena Lycia: Can I ask a quick question about Streaming costs before the end?

[13:00] Andrew Linden: Sure Helena.

[13:01] Helena Lycia: If streaming cost is linked to level of detail and sending stuff to the viewer. Would an invisible mesh count as nothing?

[13:01] Helena Lycia: Or haven't I understood streaming?

[13:01] Andrew Linden: No I don't think so.

[13:01] Kadah Coba: I think "Viewer evolution" is the closest UG for UI

[13:01] Andrew Linden: The server doesn't actually know what is invisible or not.

[13:01] Latif Khalifa: you'd assume streaming cost of regular prims have any basis in reality. it doesn't

[13:01] Andrew Linden: It just supplies the invisible texture info

[13:01] Andrew Linden: just like it would for a visible object.

[13:01] Liisa Runo: and will the cost change when people zoom their camera closer and further :P

[13:01] Vincent Nacon: don't think you can have a invisible mesh anyway (not with alpha, of course)

[13:01] Pauline Darkfury: invisible stuff still gets sent out as normal, pretty much

[13:01] Draconis Neurocam: theres no way to make something invisible in sl, it would have to not have a face, and unless you are using legacy prims i dont know of a way to do it

[13:02] Helena Lycia: SO when I just want to affect the physical shape of an object I have to pay for data sent to the viewer that isn't needed?

[13:02] Helena Lycia: You can make something 100% transparent

[13:02] Draconis Neurocam: its still rendered somewhere

[13:02] Andrew Linden: Well... I believe the streaming cost was supposed to try to balance some of the "bigger is more visible" issue.

[13:02] Vincent Nacon: just because it has transparent, doesn't meant it's not there anymore

[13:03] Andrew Linden: So, the short answer is "No", invisible objects would cost the same as similary build visible ones.

[13:03] Vincent Nacon: unless SL gave us a "culling" object

[13:03] Vincent Nacon whispers: :P

[13:03] Helena Lycia: Well I want the mesh to affect physics but it's hidden from view because it's just there to make a vehicle move proeprly

[13:03] Draconis Neurocam: its still going to add to the cost

[13:04] Draconis Neurocam: somewhere

[13:04] Vincent Nacon whispers: yup

[13:04] Leonel Iceghost: but bigger things 60m away are not visible with dd 50, and small things are visible 200 meters away with dd 200.. so I don't get it

[13:04] Pauline Darkfury: Sincere thanks for getting to 2 of my bigger issues, Andrew. Very much appreciated. (encroachment return & the sudden bloat return)

[13:04] Vincent Nacon: augh. stupid whisper key

[13:04] Helena Lycia: Well I'm trying to replace 2 complex prims with a simple mesh

[13:04] Draconis Neurocam: a mesh will never be less than two prims

[13:04] Helena Lycia: My problem isn't that it's less than 2 prims

[13:04] Latif Khalifa: Helena, stick to prims

[13:04] Vincent Nacon: stick with sculpty

[13:04] Helena Lycia: It's that is 200 times more than 2 prims

[13:05] Helena Lycia: Sculpties don't help with physics

[13:05] Vincent Nacon: use mesh for that

[13:05] Andrew Linden: Thanks for coming everyone.

[13:05] Sahkolihaa Contepomi: See you, Andrew.

[13:05] Liisa Runo: thanks everyone

[13:05] Vincent Nacon: but use sculpty for visual

[13:05] Morgaine Dinova: Thanks Andrew, tc :-)

[13:05] Helena Lycia: Can't because the PE ends up at 450 and that prevents physics

[13:05] Draconis Neurocam: thanks, simon, andrew and kelly

[13:05] Pauline Darkfury: Thanks, Lindens, have a good afternoon :)

[13:05] Latif Khalifa: take care andrew

[13:05] Helena Lycia: Thanks Andrew

[13:05] Vincent Nacon: oh well

[13:05] Vincent Nacon: thanks all

[13:06] Liisa Runo: use normal prims for collisions, sculpts for visual, and mesh for hair


Prev 2011.08.05 Next 2011.08.12