User:Andrew Linden/Office Hours/2009 04 14
Jump to navigation
Jump to search
Transcript of Andrew Linden's office hours:
[11:08] | Moon Metty: | hi Simon :) |
[11:08] | Latif Khalifa: | hi simon :) |
[11:08] | Uni Ninetails: | hihi! |
[11:08] | Helena Lycia: | Hiya Simon |
[11:08] | Simon Linden: | Hello, sorry to be late |
[11:08] | Roberto Salubrius: | morn |
[11:08] | Rex Cronon: | hi |
[11:08] | Morgaine Dinova: | Hiya Simon! |
[11:08] | Bear Hunt Start: Bear Quest Start touched by Latif Khalifa | |
[11:09] | Simon Linden: | Just spent time struggling with the viewer -- had some odd startup problem |
[11:09] | Morgaine Dinova: | My first Simon-only session :-) /me wonders if topics still roughly the same :-) |
[11:09] | Latif Khalifa: | is andrew no longer attending these? |
[11:09] | Morgaine Dinova: | He's on holiday |
[11:09] | Twisted Laws: | vacation |
[11:09] | Latif Khalifa: | bastard! |
[11:09] | Simon Linden: | Andrew is on vacation |
[11:09] | Morgaine Dinova: | lol |
[11:09] | Latif Khalifa envious | |
[11:09] | Helena Lycia: | Hey, I'm on holiday and I'm here |
[11:10] | Moon Metty: | lol |
[11:10] | Roberto Salubrius: | hol.. holiday... what's that ... I have not taken one of those in like 15 years |
[11:10] | Morgaine Dinova: | I've been on holiday for over a year ... the benefits of self-employment and only working when you need to :P |
[11:10] | Simon Linden: | Actually I'm also taking the day off, but have some time as I'm just doing errands |
[11:10] | Morgaine Dinova: | Simon: that's dedication :-) |
[11:11] | Latif Khalifa: | good excuse to skip a few errands, hon, i need to do this thing for work :P |
[11:11] | Simon Linden: | So the 1.26 server is struggling to get out -- there was another rollout scheduled for this morning to the pilot regions |
[11:12] | Simon Linden: | There were a few permissions problems that had to be fixed or rolled back |
[11:12] | Moon Metty: | Simon, the permission bugs are solved |
[11:12] | Rex Cronon: | is a new viewer equired for he new server |
[11:12] | Rex Cronon: | ? |
[11:12] | Twisted Laws: | 1.26.2 seems to be doing good so far :) |
[11:12] | Simon Linden: | Moon - yes, except for the one we were trying to fix :( |
[11:12] | Moon Metty: | really? |
[11:12] | Morgaine Dinova: | Simon, I'm involved in the PyOGP thing with Enus and Sai ... is Enus's use of PyOGP scripting helping you in those rollouts, yet? I know it's intended to help in unit and regression testing later. |
[11:12] | Moon Metty: | hmmmm |
[11:13] | Moon Metty: | which one is that then? |
[11:13] | Latif Khalifa: | they fixed one, created 4 critical ones :P |
[11:14] | Simon Linden: | The bug is one with linked objects being so0ld and then unlinked. I forget the jira |
[11:14] | Moon Metty: | is that a SEC issue? |
[11:14] | Moon Metty: | or SVC-4100 |
[11:14] | Morgaine Dinova: | Hmmm, I've just realized I don't know the relationship between Simon, Andrew, Prospero, and Enus/QA. |
[11:14] | Latif Khalifa: | no, they tried to fix some bug, and the "fix" created svc-4100 |
[11:15] | Simon Linden: | Andrew is old-time development - he's employee #2 (Philip's first hire) |
[11:15] | Morgaine Dinova: | Hehe |
[11:16] | Simon Linden: | I'm a developer and have worked with Andrew on server stuff, hence lurking and sharing office hours |
[11:16] | Simon Linden: | Prospero is in release engineering - he manages the rollouts, for example |
[11:16] | Latif Khalifa: | andrew and simon, coders, prospero and vek release team and q |
[11:16] | Latif Khalifa: | qa |
[11:17] | Simon Linden: | Enus is in QA, so he tests stuff |
[11:17] | Helena Lycia: | Does SVC-4100 bork perms the other way, could a no-trans object be made full perms? |
[11:17] | Simon Linden: | ANyway, 1.26 is looking good |
[11:17] | Latif Khalifa: | Helena SVC-4100 made permissions more restrictive, not less |
[11:17] | Helena Lycia: | Few! |
[11:18] | Moon Metty: | SVC-4100 is fixed, i tested it with Vek yesterday on aditi |
[11:18] | Morgaine Dinova: | Aha! And does QA work with developers directly (and hence relevant to you), or is your QA post-development at release time (and hence more a matter for Prospero)? |
[11:19] | Simon Linden: | In general you just need to be careful with perms before linking if your going to sell an object someone can unlink |
[11:19] | Simon Linden: | Morgaine - there's both. We'll test during development, and of course before releasing |
[11:20] | Eata Kitty: | I would blame this problem on bad interface design Ie lack of mass perms editing |
[11:20] | Moon Metty: | 1.23 will have mass perm setting |
[11:20] | Eata Kitty: | Yes, five years lateish |
[11:21] | Simon Linden: | Yes, and the rules on what should happen to embedded or linkied objects when sold are fuzzy |
[11:21] | Moon Metty: | yeah .. permissions are tough |
[11:22] | Morgaine Dinova: | Simon: devs testing and own QA is assumed, hehe. But I was wondering whether Enus' work is with yourselves directly, or just on release QA. (The question is because of our PyOGP stuff --- wondering if it'll be mostly for release QA) |
[11:23] | Morgaine Dinova: | Aye, perms are very complex. And interop will make them even more so. |
[11:23] | Simon Linden: | Well, we want to have more automated QA available, which is useful all over |
[11:23] | Morgaine Dinova nods | |
[11:24] | Simon Linden: | It keeps us from introducing rgression bugs |
[11:24] | Morgaine Dinova: | Aye |
[11:24] | Latif Khalifa: | lola do some automated stuff lol |
[11:24] | Morgaine Dinova: | Crucial for an evolving system. |
[11:24] | LoLa Varriale: | I do nothing but chat all day.Quite honestly, I wouldn't worry myself about that.I'm glad you find this amusing. |
[11:24] | Moon Metty: | lola be quiet |
[11:24] | Simon Linden: | And if we can get a good test system up and running, it's easier to add tests for new features |
[11:24] | LoLa Varriale: | As far as I know, I am not making any sound at all. |
[11:24] | Morgaine Dinova: | Lola: go program something |
[11:24] | LoLa Varriale: | Where is that? |
[11:25] | Latif Khalifa: | lol |
[11:25] | Morgaine Dinova: | Simon: thanks. Won't ask more about QA, back to perms etc :-) |
[11:26] | Simon Linden: | I was looking at some of the AI / XMl stuff the other day. Interesting work there |
[11:26] | Morgaine Dinova: | Linkie? |
[11:26] | Latif Khalifa: | yeah, lol@ uses aiml |
[11:27] | Latif Khalifa: | she is parsing chat, and she hears her name in it, she'll try answering hehe |
[11:27] | Morgaine Dinova: | http://en.wikipedia.org/wiki/AIML |
[11:27] | Simon Linden: | right, aiml. I'm a newbie to AI but find bots fascinating |
[11:27] | Moon Metty: | yes, it is :) |
[11:28] | Simon Linden: | I'm going to re-log : I'm on a really slow netbook and want to get on a faster machine |
[11:30] | Moon Metty: | welcome back Simon |
[11:30] | Eata Kitty: | God earlier I thought Latif was some kind of evil slave owner who was abusing lola but I didnt comment on it. Good thing huh |
[11:30] | Latif Khalifa: | wb Simon :) |
[11:30] | LoLa Varriale: | When was this exactly? Yup. It is good, isn't it. |
[11:30] | Helena Lycia: | wb |
[11:30] | Uni Ninetails: | Lol what is the maxiumum airspeed of an unladen swallow? |
[11:30] | Rex Cronon: | wb |
[11:31] | Uni Ninetails: | bugger |
[11:31] | Rex Cronon: | . |
[11:31] | Latif Khalifa: | Eata, lol, no I was just chatting with an AI |
[11:31] | Latif Khalifa: | she gets defensive lol |
[11:31] | Morgaine Dinova: | Latif: I liked the av you gave her once when at Prospero's OH on Beta Grid, less in-your-face :-) |
[11:31] | Latif Khalifa: | if I call her stupid |
[11:32] | Morgaine Dinova: | Wb Simon :-) |
[11:32] | Latif Khalifa: | Morg, my gf dresses her up, I cannot be bothered |
[11:32] | Morgaine Dinova: | Rex: sticky keyboard? Should eat less jam :-) |
[11:32] | Rex Cronon: | why am i stuck in typing? |
[11:32] | Eata Kitty: | Rex: Sl gets stuck on a lot of things which I blame on UDP, probably out of sequence messages |
[11:33] | Simon Linden: | So is Lo La built on pyogp? |
[11:33] | Latif Khalifa: | no |
[11:33] | Morgaine Dinova: | Latif: ah, hehe. She's a good dresser. |
[11:33] | Latif Khalifa: | libomv, pygop i way off |
[11:33] | Latif Khalifa: | not even close to this level |
[11:33] | Morgaine Dinova: | Yes, PyOGP isn't even close yet. |
[11:33] | Rex Cronon: | nothing sticky. laggy yes:( |
[11:34] | Simon Linden: | Her client then just runs a console window? I'm not very familiar with the different SL libraries |
[11:34] | Latif Khalifa: | if you know what mirc looks like, something like that |
[11:34] | Morgaine Dinova: | kk |
[11:34] | Latif Khalifa: | yes, its a text-only client |
[11:34] | Helena Lycia: | I think the true test of a bot in SL won'tbe emulating talking like a real person but shopping like a real person |
[11:34] | Simon Linden: | ok |
[11:35] | Latif Khalifa: | but it supports most of sl functionality, short of rendering the 3d scene |
[11:35] | Morgaine Dinova: | Latif: 2D radar? |
[11:35] | Latif Khalifa: | yeah she has a sim scanner |
[11:35] | Simon Linden: | I'd love to have python embedded in the viewer so you could get the 3d scene as well |
[11:35] | Morgaine Dinova: | Super! |
[11:36] | Latif Khalifa: | yeah that would be very nice :) |
[11:36] | Simon Linden: | That would be useful for us to drive tests. I know someone internally did it, but got stuck on the object/interface model |
[11:36] | Latif Khalifa: | problem with sl viewer that its not really componentised |
[11:36] | Latif Khalifa: | just the other day i discovered that there is no api call to send chat for example |
[11:37] | Morgaine Dinova: | Really? |
[11:37] | Latif Khalifa: | the gui chat component composes and sends the packet directly |
[11:37] | Morgaine Dinova: | Ouch |
[11:37] | Simon Linden: | Right - it's a lot of work to design the object model for the language, and then wiring it up internally is really ugly with the current viewer code |
[11:37] | Latif Khalifa: | i was expecting to have something like gAgent.sendChat() |
[11:37] | Latif Khalifa: | but to my horror found that the chat bar gui element sends the wire packet directly |
[11:37] | Morgaine Dinova: | Pity Aimee isn't here, I'd like her comments on that. |
[11:38] | Latif Khalifa: | yeah, the viewer would need a generic packet/network module that would dispatch messeges around to subsribers |
[11:38] | Simon Linden: | yeah, SL has unfortunately evolved too fast without some of the slower design work. |
[11:39] | Morgaine Dinova: | And nobody in SLdev is interested in refactoring? |
[11:39] | Eata Kitty: | Just a slightly large task I think... |
[11:39] | Morgaine Dinova: | Ah, I know the problem. Without LL's blessing, such refactoring by the community would get lost in the next LL release. |
[11:40] | Simon Linden: | We need to fix the way SL dev works first ... the path to get code merged is too long and difficult for a big project like that |
[11:40] | Latif Khalifa: | I have seen a patch already that does this. Generic event dispatcher |
[11:40] | Simon Linden: | That's another long discussion, however |
[11:40] | Morgaine Dinova: | Either LL buys into the community process, or there is no community refacgoring. |
[11:40] | Morgaine Dinova: | Simon: agreed |
[11:40] | Simon Linden: | Agreed ... there's been some discussion internally about that, so I think it will get better |
[11:40] | Rex Cronon: | i got disconnected:( |
[11:40] | Morgaine Dinova: | Rex: it's that jam |
[11:41] | Moon Metty: | lol |
[11:41] | Morgaine Dinova: | It's go into Rex's modem. |
[11:41] | Rex Cronon: | modem ok |
[11:41] | Simon Linden: | Let's see ... I have a side project going on that should be interesting to builders. It's to allow setting some parameters to determine the shape used for an object in the physics engine |
[11:41] | Rex Cronon: | didn't lost connection to the net. only sl view logged me off |
[11:42] | Morgaine Dinova: | Simon: WOOOOT!!!!! User-defined bounding boxes? |
[11:42] | Moon Metty: | interesting Simon |
[11:42] | Simon Linden: | So you might have a complex looking object, but can simulate the shape with a sphere or box |
[11:42] | Rex Cronon: | only bounding boxes:( |
[11:42] | Morgaine Dinova: | Very cool Simon. I think that's necessary |
[11:43] | Morgaine Dinova: | Simon: oh, you mean no complex BBs? |
[11:43] | Latif Khalifa: | that would be cool. so you could have vehicles with more than 31 prims |
[11:43] | Rex Cronon: | could we give an custome object to havok |
[11:43] | Rex Cronon: | ? |
[11:43] | Helena Lycia: | Cool |
[11:43] | Helena Lycia: | Very cool |
[11:43] | Simon Linden: | It's evolving ... right now you can pick the shape to use for a prim. Choices are simple - sphere, box, full detail (the current mode) or none (if it's linked) |
[11:44] | Morgaine Dinova: | Ah. Well better than nothing, but really needs those basic primitives to be composable into arbitrary BB too |
[11:44] | Simon Linden: | I used it on the kart in the library, and reduced it to 1/3 the complexity. Made the sim much faster - I got almost a 3x performance boost, but there was nothing but karts on the sim |
[11:44] | Morgaine Dinova: | Hehe |
[11:44] | Helena Lycia: | Any idea when it will be available? |
[11:44] | Simon Linden: | We're looking into being able to replace the shape for an entire linked object with a bounding box or sphere, possibly with some scaling or offset |
[11:45] | Latif Khalifa: | Simon, how many devs do develop the simulator code? |
[11:45] | Simon Linden: | No idea on the timing ... it's still experimental. I'm working with an intern on it, so hopefully sometime later this year |
[11:45] | Eata Kitty: | I know a ton of things could just be a box and that would give pretty decent results |
[11:45] | Helena Lycia: | Any other shapes, boxes don't seem to do well over rough ground and spheres tend to fall over |
[11:45] | Simon Linden: | yeah, the current design uses way more complexity than needed in a lot of cases |
[11:46] | Morgaine Dinova: | Eeeek! That's heading the WRONG way, if you mean that the BBs of composite objects currently will be replaced by *less detailed* BBs. |
[11:46] | Latif Khalifa: | yeah, most vehicles don't need correct collisions down to the last prim |
[11:46] | Eata Kitty: | Why not Morgaine? You only need very rough collision model for almost anything |
[11:46] | Simon Linden: | There still some questions ... like if you had two linked objects that aren't touching, we probably don't want to use a big box around the whole thing |
[11:46] | Eata Kitty: | A couple of boxes nicely cover the collision shape for all kinds of things |
[11:47] | Helena Lycia: | A vehicle I work on seems to rely on some unusual properties of its physical form |
[11:47] | Simon Linden: | Morgaine - it's totally a builder's option. The default would be to work the way it currently does |
[11:47] | Twisted Laws: | issue of things like trains that need to straddle that guide on the linden railroads :p |
[11:47] | Latif Khalifa: | Simon, if you want a project in the spear time :) Most people would *love* is sculpties collided correctly -- so you could walk on sculpted staris fx |
[11:48] | Morgaine Dinova: | Eata: it depends on the case. In the case of composite constructions where the inner BBs aren't reachable, sure, it makes sense to simplify the outer BB. But when a surface is reachable the BB should conform to it in general. |
[11:48] | Simon Linden: | Yes, fixing sculpties in physics would be nice. This is somewhat related, so maybe a part of this work will help out |
[11:48] | Morgaine Dinova: | Simon: cool! |
[11:48] | Moon Metty: | the havok engine doesn't know about the shape of a sculptie |
[11:48] | Latif Khalifa: | Moon, currently no, but that's up to sim code to tell the physics engine |
[11:49] | Rex Cronon: | yes. sculpties;) |
[11:49] | Moon Metty: | yes :) |
[11:49] | Simon Linden: | There are others thinking about meshes in general ... it would be really nice to open up mesh support, importing models, etc., but that's getting into a lot of code |
[11:49] | Latif Khalifa: | they got that part working in opensim |
[11:49] | Simon Linden: | As usual, we all have too many ideas and not enough time :) |
[11:49] | Latif Khalifa: | hehe the eternal time problem :P |
[11:50] | Moon Metty: | well, stability is much much better now, but there are still a lot of bugs to be fixed |
[11:50] | Morgaine Dinova: | Simon: re meshes ... indeed. And from my interop/MMOX angle, that's not optional --- there won't be interop without mesh import (including for avs), because 95% of the rest of the graphics world uses meshes. |
[11:51] | Latif Khalifa: | morg, i disagree, SL's prim model is briliant for transmiting live scene over the net |
[11:51] | Simon Linden: | Hopefully we haven't introduced any new ones :) |
[11:51] | Simon Linden: | We're hoping 1.26 will get even better with the server crash rate ... we fixed our most common crash mode, which was something like 30% or more of the crashes |
[11:51] | Simon Linden: | I'm getting some odd chat delays, fwiw. Those last two messages were out of order |
[11:51] | Eata Kitty: | UDP noooo! |
[11:52] | Latif Khalifa: | lol another udp hater |
[11:52] | Latif Khalifa: | if this was tcp, you would not be getting any messages at all |
[11:52] | Morgaine Dinova: | Simon: factor out small independent pieces of the system, set up a community project for it, and let FOSS add their manpower. It means loss of control, but you have no option really, unless you intend to set up a Microsoft-size dev team. |
[11:52] | Latif Khalifa: | waiting for the tcp to get unstuck retransmitting |
[11:52] | Simon Linden: | Put me in that crowd -- it causes all sorts of problems when you try an use it for something reliable |
[11:52] | Eata Kitty: | Unreliable Datagram Protocol |
[11:53] | Simon Linden: | Looks like we lost Rex again |
[11:53] | Helena Lycia: | How careless of us |
[11:53] | Moon Metty: | maybe it's peanutbutter instead of jam |
[11:54] | Morgaine Dinova: | Latif: that wasn't the point. Yes, prims are great. But for interop, you're not going to convert the remainder of the world to abandon their meshes. Interop is about letting each VW stick to their own designs, and port them to at least some degree. |
[11:54] | Latif Khalifa: | interop with the rest of the world is a pipe dream :P |
[11:55] | Morgaine Dinova: | Some people still have dreams, foretunately :-) |
[11:55] | Twisted Laws: | it will happen sometime, somewhere |
[11:55] | Helena Lycia: | wb Rex |
[11:55] | Twisted Laws: | what we have now used to only be a dream |
[11:55] | Morgaine Dinova: | Twisted++ |
[11:55] | Rex Cronon: | thanks |
[11:55] | Helena Lycia: | You have some weird dreams :) |
[11:55] | Rex Cronon: | something is wrong here:( |
[11:55] | Simon Linden: | I think there's always going to be some conversion/import/export when jumping between systems. |
[11:56] | Morgaine Dinova: | Those who call things "pipe dreams" are merely saying how close their horizons are. |
[11:56] | Simon Linden: | Someone else is always going to add some feature to their VW that isn't supported in others |
[11:56] | Latif Khalifa: | sure sure keep dreaming. I am still waiting for LL to opensource the simulator code which they promissed to do two years ago :D |
[11:56] | Morgaine Dinova: | Yep |
[11:56] | Eata Kitty: | Depending on feature set I dotn think meshes are going to be any better off for import/export than prims :| |
[11:56] | Arawn Spitteler remembers the time, just in time, to mention SVC-22, SVC-2931 and SVC-93 | |
[11:56] | Arawn Spitteler: | And Sheep |
[11:56] | Simon Linden: | Right, it all comes down to the feature set. |
[11:57] | Latif Khalifa: | you think it takes too long to download a scene now? wait until you have to wait for all the meshes to downoad :P |
[11:57] | Morgaine Dinova: | Interop isn't about open-sourcing VW code. It's about interworking between VWs. Whether they're open or close source is not part of the interop discussion. |
[11:57] | Simon Linden: | Also 3D modelers often work at cross purposes - super high detail for beautiful models means a horrible real-time experience |
[11:57] | Eata Kitty: | Personally if I was LL I would never open source my sim |
[11:58] | Morgaine Dinova: | Rex: big troubles? |
[11:58] | Latif Khalifa: | Eata, point is they did say they were going to... |
[11:58] | Simon Linden: | I haven't heard much discussion heading towards that lately, although I probably wouldn't be in the conversation anyway |
[11:58] | Eata Kitty: | Well I wouldn't expect them to |
[11:58] | Arawn Spitteler: | Refusal to opensource leads to a more creative environment. Viewer is already opensource, so different environemnts could be interfaced. |
[11:58] | Eata Kitty: | Besides I imagine its incredibly specialised and would be extremely hard to use |
[11:59] | Latif Khalifa: | LL does offer SL-in-a-box now, sim code and all, for "enterprise customers" |
[11:59] | Morgaine Dinova: | I hope LL doesn't open-source the sim code, because it would hardwire the world into non-scalable regions. |
[11:59] | Eata Kitty: | Is the custom grid one still running on LL servers or actual company owned hardware in their place |
[11:59] | Morgaine Dinova: | As things stand, the open community will make better region implementations, so we all benefit (except LL). |
[12:00] | Simon Linden: | That's one huge mountain I wish we'd try to climb ... more choice in region size would be really, really nice |
[12:00] | Rex Cronon: | i think i know what is wrong. there r over 10k objects in the sim. whenever i get to a sim like this lag is terrible:( |
[12:00] | Eata Kitty: | I heard your maximum region size isn't properly defined as a constant :P |
[12:00] | Arawn Spitteler: | I think the Grid as Product would have to retain map sizes, but how big a meter might be called into question |
[12:01] | Morgaine Dinova: | Simon: it's not about acreage, but about scalability. Your sim code **for a single region** can't run across a cluster of machines. |
[12:01] | Twisted Laws: | shorten your draw distance at meetings Rex |
[12:01] | Eata Kitty: | Oh I dont know rex, your client only gets told about a small amount relative to you so potentially a thousand objects in one space could be as bad as 10k in a sim |
[12:01] | Simon Linden: | Well, there are a lot of ways to scale. If we just had 1km x 1km regions, for example, you could make more fun sailing, flying or racing sims by avoiding region crossings |
[12:02] | Uni Ninetails: | agreed |
[12:02] | Helena Lycia: | Please!!!! |
[12:02] | Moon Metty: | indeed |
[12:02] | Eata Kitty: | I would bet virtually everyone would increase their region size if it was an option |
[12:02] | Helena Lycia: | Sim crossings have been horrible for about a week now |
[12:02] | Arawn Spitteler: | Vary the meters per simlength? |
[12:02] | Uni Ninetails: | for the loave of god please 1km x 1km |
[12:02] | Moon Metty: | 1 ly x 1 ly |
[12:03] | Simon Linden: | Sim crossings should be a bit better in 1.26 -- the code that parses objects entering the new region is a bit faster, so hopefully that will help |
[12:03] | Arawn Spitteler: | Changing the Meters per sim-long might make for interesting sim crossings, but should be doable. |
[12:03] | Morgaine Dinova: | Simon: Zero has actually stated during an Office Hours that he (or LL, can't remember words) has no interest in making regions scalable, and that therefore it would be up to third-party implementors to make scalable regions if they need them. |
[12:03] | Latif Khalifa: | i would not hold my breath lol... sim size of 256m is soo at the core of many things--- untangling it would be next to impossible :D |
[12:03] | Helena Lycia: | I'd be happy with 1km x 1km, hell I'd be happy with 512m x512m |
[12:03] | Latif Khalifa: | i happen to know how global coordinates gets translated to local |
[12:03] | Helena Lycia: | We can dream can't we? |
[12:03] | Latif Khalifa: | lol |
[12:03] | Simon Linden: | The viewer and LSL actually are problems ... there are too many places that just expect another region to be there when you move more than 256m |
[12:04] | Arawn Spitteler: | The Viewer is open source. |
[12:04] | Morgaine Dinova: | Scalability is not about acreage. |
[12:04] | Rex Cronon: | 64 m draw distance |
[12:04] | Latif Khalifa: | you get global coordinate x, y, so which map position of the sim it is. answer x/256 and y/256, a lot of code like that |
[12:05] | Simon Linden: | Right, I did a search for 256 in our codebase and was horrified to see how many places it's just hard-coded :( |
[12:05] | Moon Metty: | lol |
[12:06] | Latif Khalifa: | i've seen it in these local <-> global coordinate translations, making different size regions next to impossible |
[12:06] | Simon Linden: | It's not even easy to change that number, re-build the sim and viewer, and get it to work. |
[12:06] | Latif Khalifa: | short of a complete rewrite :P |
[12:06] | Simon Linden: | I'm going to have to go soon ... any last questions? |
[12:07] | Helena Lycia: | If a sim can't be upscaled then develop a way that multiple sims geographically connected can run on the same server and have improved comms between them so sim crossings between sims on the same server have negligible impact on use experience |
[12:07] | Morgaine Dinova: | We actually started off a Scalability Group in the AWG, way back. It all fell apart when it became clear that Zero wasn't interested. -- https://wiki.secondlife.com/wiki/Scalability_VAG |
[12:07] | Simon Linden: | If you see any issues with 1.26, please let support know ASAP and file a jira. We're hoping to roll it out very soon, but want to catch any bugs we can |
[12:07] | Moon Metty: | thank you for spending time with us Simon :) |
[12:07] | Morgaine Dinova: | Thanks Simon :-) |
[12:07] | Eata Kitty: | Yes, thankyou |
[12:07] | Helena Lycia: | Yes, thanks Simon |
[12:08] | Uni Ninetails: | tyty :) |
[12:08] | Latif Khalifa: | yes, thanks for coming on your day off :)) |
[12:08] | Twisted Laws: | thanks :) |
[12:08] | Simon Linden: | Thanks everyone for coming ... we'll have Thursday's office hours at the usual time |
[12:08] | Morgaine Dinova: | Super :-) |
[12:08] | Simon Linden: | I'm not sure if Andrew is back then or not, but I'll be here |
[12:08] | Arawn Spitteler mumbles something about Sheepo and SVC-22, SVC-93 and SVC-2931: Thanks for coming, Simon | |
[12:08] | Moon Metty: | very nice |
[12:09] | Arawn Spitteler: | Isn't Andrew deployed to some personal time? |
[12:09] | Simon Linden: | yeah, we gotta sort out the rules about vehicles. That's good for another day |
[12:09] | Arawn Spitteler: | "Recreatiuonal Individuality? |
[12:09] | Simon Linden: | yep, Andrew's on well deserved vacation |
[12:09] | Simon Linden: | OK, gotta run. Bye all! |
[12:09] | Twisted Laws: | bye |
[12:09] | Moon Metty: | see you soon! |
[12:09] | Morgaine Dinova waves | |
[12:09] | Qie Niangao: | thanks Simon |
[12:09] | Uni Ninetails: | Cya! |