User:Andrew Linden/Office Hours/2009 09 08

From Second Life Wiki
Jump to navigation Jump to search
  • [2009/09/08 11:00] GG Office Hours HUD v1.5: 11:00 AM: Andrew Linden: Technical issues, SL architecture, physics engine
  • http://slurl.com/secondlife/Denby/213/45/34 (Starts now)
  • [2009/09/08 11:00] Techwolf Lupindo: Hi andrew
  • [2009/09/08 11:00] Andrew Linden: Hello
  • [2009/09/08 11:01] xstorm Radek: OMG!!
  • [2009/09/08 11:01] Andrew Linden: hehe, what is so OMG?
  • [2009/09/08 11:01] Rex Cronon: hello everybody
  • [2009/09/08 11:01] Melchizedek Blauvelt: heya Rex
  • [2009/09/08 11:01] Rex Cronon: hi
  • [2009/09/08 11:01] xstorm Radek: i was watching kelly lindens video he has running on his land
  • [2009/09/08 11:02] xstorm Radek: itd do you wish to date my avatar *GIGGLES* :)~[2009/09/08 11:02] Andrew Linden: Oh really? What is it about?
  • [2009/09/08 11:02] Techwolf Lupindo: !!, I keep hearing "Date my avatar" latly...must be a fad.
  • [2009/09/08 11:02] Andrew Linden: Ah the music video.
  • [2009/09/08 11:02] xstorm Radek: a youtube music video
  • [2009/09/08 11:03] xstorm Radek: yes a bit funny but made me thing there is more people can do with that
  • [2009/09/08 11:03] Andrew Linden: Ok Simon Linden says he'll be a little late but will show.
  • [2009/09/08 11:03] Techwolf Lupindo: Simon said
  • [2009/09/08 11:03] Techwolf Lupindo: :-)
  • [2009/09/08 11:03] Melchizedek Blauvelt: sounds like it's an Aphex Twin vid heh
  • [2009/09/08 11:03] Andrew Linden: Looks like server-1.30 has been postponed one week.
  • [2009/09/08 11:03] Techwolf Lupindo: arg...what was the child game...simon says I think.
  • [2009/09/08 11:04] Andrew Linden: So the deploy should now start around Wed Sept 16, instead of the 9th
  • [2009/09/08 11:04] xstorm Radek: what if there was a real life feed from a cam system into SL on a scaled wall thats our scale
  • [2009/09/08 11:04] Liisa Runo: will dialog spam be fixed in 1.3 ?
  • [2009/09/08 11:04] Melchizedek Blauvelt: It is, Tech, never watched Demolition Man I suppose
  • [2009/09/08 11:04] Andrew Linden: xstorm, that kind of mixed reality setup has been done a few times
  • [2009/09/08 11:04] GG Office Hours HUD v1.5: 11:00 AM: Lexie Linden: SL Volunteering
  • http://slurl.com/secondlife/Ambleside/181/99/30 (Started 5 minutes ago)
  • [2009/09/08 11:04] GG Office Hours HUD v1.5: 11:00 AM: Andrew Linden: Technical issues, SL architecture, physics engine
  • http://slurl.com/secondlife/Denby/213/45/34 (Started 5 minutes ago)
  • [2009/09/08 11:05] Techwolf Lupindo: Demolition Man been a while ago.
  • [2009/09/08 11:05] Liisa Runo: i ask this cause i thought dialog spam been fixed already, but today i was proven wrong
  • [2009/09/08 11:05] xstorm Radek: yes but grab the textures of real people to put on the avatars
  • [2009/09/08 11:05] Andrew Linden: which dialog spam Liisa?
  • [2009/09/08 11:05] Liisa Runo: object opens 1000000000000000000 dialogs
  • [2009/09/08 11:06] Liisa Runo: you dont crahs anymore, but relog is still the only way to get rid of them
  • [2009/09/08 11:06] Techwolf Lupindo: I need to test taht again. But the test I ran exposed a coupld quirks that that Blue Dialog attack.
  • [2009/09/08 11:06] xstorm Radek: a video image can be grab
  • [2009/09/08 11:06] Liisa Runo: crash*
  • [2009/09/08 11:06] Andrew Linden: not that I know of Liisa. Does anyone have a jira number for that dialog spam problem?
  • [2009/09/08 11:06] Techwolf Lupindo: I have an oject I created for testing. I click on it and the person that click on it will get Blue Dialongs at a rate of 100 per secnd.
  • [2009/09/08 11:07] Techwolf Lupindo: Will drop frame rate to less then one on most clients.
  • [2009/09/08 11:07] Simon Linden: Hmm, there should be a throttle on that, but maybe it's not working or the object found a way around it
  • [2009/09/08 11:07] Liisa Runo: no idea of a jira, i just experienced the attack 30min ago, and before that i thought it been fixed
  • [2009/09/08 11:07] Techwolf Lupindo: emerald has a patch to prevent that.
  • [2009/09/08 11:07] Andrew Linden: there are throttles on some of those alerts, I think
  • [2009/09/08 11:07] Techwolf Lupindo: emerald is immune to the Blue Dialog attack.
  • [2009/09/08 11:08] Andrew Linden: Techwolf, I'd like a copy of that object if you're willing to hand it over.
  • [2009/09/08 11:08] Liisa Runo: and yea, alternative viewers have had fix for over an year now
  • [2009/09/08 11:08] Rex Cronon: i think that attack doesn't work if u mute theowner
  • [2009/09/08 11:08] xstorm Radek: throttling got to be so bad one day when i was in corfeld that people got stuck and crashed
  • [2009/09/08 11:09] Andrew Linden: well, any trigger for server transmission should probably be throttled
  • [2009/09/08 11:09] Andrew Linden: so I'm interested in fixing it server-side too, despite client capabilities
  • [2009/09/08 11:09] Liisa Runo: and maybe some kind of "close all dialogs" -button to viewer, or something
  • [2009/09/08 11:09] Rex Cronon: test.
  • [2009/09/08 11:10] Techwolf Lupindo: Let me see if I can find it. I created it months ago.
  • [2009/09/08 11:10] Andrew Linden: I think I mentioned in another office hour, but I should probably repeat:
  • [2009/09/08 11:11] Andrew Linden: Linden Lab is currently in a hiring sprint and there are more than the usual number of job openings available
  • [2009/09/08 11:11] Andrew Linden: in particular we're looking for some C++/Python devs who want to work on the server backend
  • [2009/09/08 11:11] Rex Cronon: sorry. my text arrives in the worng orer
  • [2009/09/08 11:11] Techwolf Lupindo checks the employment page...
  • [2009/09/08 11:11] Andrew Linden: the positions are listed here:
  • [2009/09/08 11:12] Andrew Linden: http://lindenlab.hrmdirect.com/employment/openings.php?sort=da&
  • [2009/09/08 11:12] Andrew Linden: Lesssee... what I'm working on this week...
  • [2009/09/08 11:12] Techwolf Lupindo: arg...why do I have to log out to get the secondlife page?
  • [2009/09/08 11:13] xstorm Radek: ?
  • [2009/09/08 11:13] Arawn Spitteler: Sounds like a client side normality
  • [2009/09/08 11:13] Andrew Linden: I've got some work to do on internal tools for content management to do this week
  • [2009/09/08 11:13] Techwolf Lupindo notices the word "senior" in from of the job titles.
  • [2009/09/08 11:13] Techwolf Lupindo: front
  • [2009/09/08 11:13] xstorm Radek: Techwolf thats not the norm
  • [2009/09/08 11:13] Andrew Linden: relating to copyright and trademark enforcement
  • [2009/09/08 11:14] Aimee Trescothick: just dye your fur grey Tech
  • [2009/09/08 11:14] Andrew Linden: other than that I'm going to be fixing bugs that I hope to get into server-1.32
  • [2009/09/08 11:14] xstorm Radek: *GIGGLES* :)~[2009/09/08 11:15] Andrew Linden: yeah Techwolf, I actually raised that issue with the LL dev who was writing those job requirements...
  • [2009/09/08 11:15] Arawn Spitteler: Are we still unable to update our inventoried projects? We can't Save To Inventory? It's not like I can build Noah's Ark in a single session
  • [2009/09/08 11:15] Andrew Linden: do we only want "senior" devs? Don't we have some junior dev positions open?
  • [2009/09/08 11:15] Techwolf Lupindo: To get "sinior" devolopers, have to shell out $$$$$$
  • [2009/09/08 11:15] Andrew Linden: The response I got was, "Well that senior requirement won't prevent any junior devs from applying." But I disagree.
  • [2009/09/08 11:16] xstorm Radek: *GIGGLES* :)~[2009/09/08 11:16] Aimee Trescothick: totally disagree
  • [2009/09/08 11:16] Arawn Spitteler: It'll prevent the honest junior developers, from applying, and nobody needs those, but QA
  • [2009/09/08 11:16] xstorm Radek: thats very wrong thinking
  • [2009/09/08 11:16] Aimee Trescothick: yeah
  • [2009/09/08 11:16] Simon Linden: I know the release engineer positions are considering all levels ... gotta be a hard-core linux/python person, however
  • [2009/09/08 11:16] xstorm Radek: that be a big lag of hands on skills and training
  • [2009/09/08 11:16] Techwolf Lupindo: I even tried the QA stuff, got no responce.
  • [2009/09/08 11:16] Andrew Linden: He also stated that we'll probably open some junior dev job requirements soon, but that we want some devs that can get up to speed pretty quick right now.
  • [2009/09/08 11:17] Morgaine Dinova waves quietly
  • [2009/09/08 11:17] xstorm Radek: hi morgaine
  • [2009/09/08 11:17] Rex Cronon: hi
  • [2009/09/08 11:17] Arawn Spitteler: Why not just post what they'll be doing on the White Board?
  • [2009/09/08 11:17] Teravus Ousley:
  • [2009/09/08 11:17] Morgaine Dinova: Hi xstorm
  • [2009/09/08 11:18] Arawn Spitteler: Better, interview them in Advanced Scripters Chat
  • [2009/09/08 11:19] Techwolf Lupindo: Been thinking of joining that groups, even though I don't do LSL scripting all the time like content creatores.
  • [2009/09/08 11:19] Arawn Spitteler: Ever read P{arkinson's Law?
  • [2009/09/08 11:19] Teravus Ousley: http://www.youtube.com/watch?v=zftuoMO_R1U <--- 512x512 region code I worked on :). I turns out that there are a lot of protocol limitations that ensure that regions stay 256m^2 that one needs to 'get around' with the viewer.
  • [2009/09/08 11:19] Arawn Spitteler: Advanced Scripters of Second Life is for the imported programmers
  • [2009/09/08 11:19] Techwolf Lupindo: The junirs that apply to those are the "shotgun" types that may not be skilled enought.
  • [2009/09/08 11:20] Techwolf Lupindo: "shotgun" meaning shoot an app to anything with "computer" in it.
  • [2009/09/08 11:20] Aimee Trescothick: that's great work Teravus
  • [2009/09/08 11:20] Morgaine Dinova: Does Aditi+Vaak fall under your team area too Andrew? (Reason for question is that OGP activity is starting again, they'll be used in testing)
  • [2009/09/08 11:20] Andrew Linden: Right Techwolf. I listed half a dozen LL devs we hired that were junior, and who I'd love to clone if it were possible.
  • [2009/09/08 11:20] Aimee Trescothick: soooo wish I had regions that size for flying/driving SL vehicles in
  • [2009/09/08 11:21] Arawn Spitteler: How about an in world programmers meat market?
  • [2009/09/08 11:21] Andrew Linden: Unfortunately the current list of job requirements dissuade many smart and creative devs from applying.
  • [2009/09/08 11:21] Teravus Ousley: Adam wrote a blog entry about it : http://www.adamfrisby.com/blog/2009/09/opensim-megaregions/
  • [2009/09/08 11:21] Techwolf Lupindo: My C++ skills are begginner. But have a lot of skills doing scripting and codeing in many languages. Picking up a new one is easy for me.
  • [2009/09/08 11:21] Arawn Spitteler: Inux Python? Isn't Linux getting close to Kernl?
  • [2009/09/08 11:21] Andrew Linden: Morgaine, no I don't manage aditi or vaak grids.
  • [2009/09/08 11:22] Teravus Ousley: It's basically a 1 region that treats the viewer into thinking it's many regions.
  • [2009/09/08 11:22] Morgaine Dinova: Tnx for link Tera
  • [2009/09/08 11:22] Andrew Linden: What version of servers are actually deployed to those test grids can change
  • [2009/09/08 11:22] Andrew Linden: although, some grids are devoted to certain projects long term.
  • [2009/09/08 11:22] Andrew Linden: Aditi is the main "preview grid"
  • [2009/09/08 11:22] Ardy Lay: Hmm... Since LL runs 4 SL regions on a host, that could be fun.
  • [2009/09/08 11:22] Morgaine Dinova: Andrew: kk, tnx
  • [2009/09/08 11:23] Andrew Linden: so it usually has either the current server on the main grid, with some older versions
  • [2009/09/08 11:23] Andrew Linden: or else the next server when it is ready for preview, with some older versions running part of the world for comparison
  • [2009/09/08 11:23] Ardy Lay: If those 4 were adjacent and handoffs were done "internally".........
  • [2009/09/08 11:24] Andrew Linden: I think vaak is one of those devoted grids, but I'm not sure what project it is tied to
  • [2009/09/08 11:24] Techwolf Lupindo: megaregines would be good for racers. They need lots of space, but very few prims. A 1024x1024 with 5,000 prims limit is perfect for them.
  • [2009/09/08 11:25] Andrew Linden: Ah yes Teravus, we've talked about that idea in the past. A large region that communicates like several distinct ones.
  • [2009/09/08 11:25] Rex Cronon: or for war games:)
  • [2009/09/08 11:25] Andrew Linden: In my mind that was the easiest way to do it.
  • [2009/09/08 11:25] Arawn Spitteler: 1,024X256 with 15,000 would be similar to a Null Region
  • [2009/09/08 11:25] Techwolf Lupindo: Question, is there still a crossing at 256 or was it just expanded into 512 from 256?
  • [2009/09/08 11:25] Teravus Ousley:
  • [2009/09/08 11:25] Teravus Ousley: There is no actual crossing
  • [2009/09/08 11:25] Ardy Lay: Andrew, if that works then hardware selection could open up some.
  • [2009/09/08 11:26] Simon Linden: There'd be some sort of hand-off with the viewer, however, as it expects something to happen at 256. Hopefully it would be almost no overhead
  • [2009/09/08 11:26] xstorm Radek: Andrew that will fix the problem with places like korea and the griefers with the rezer prims
  • [2009/09/08 11:27] Teravus Ousley: The viewer 'looks' like you're crossing.. in that you can see the border but you're not actually crossing any border.. and as you continue to move East, for example,.. your coordinates continue to go up
  • [2009/09/08 11:27] Simon Linden: The real problem with our current crossings is stopping and serializing all the info of your AV, all scripts, etc, sending it to the next region, deserializing, and adding it to the new simulator
  • [2009/09/08 11:27] Simon Linden: That would go away on a mega-region's internal 256 boundaries
  • [2009/09/08 11:27] Techwolf Lupindo: When my viewer does the "powerwalk/fly", I often end up at 300 or greater.
  • [2009/09/08 11:27] Rex Cronon: what r talkig about xstorm?
  • [2009/09/08 11:28] Teravus Ousley: The viewer doesn't ever disconnect from the region that you're in.. technically, regardless of where you are on the MegaRegion, the viewer thinks you're in the south west most region .. as far as 'regionhandle'
  • [2009/09/08 11:28] Simon Linden: The viewer's just interpolating your last motion while the sim gets slow in doing that crossing
  • [2009/09/08 11:28] xstorm Radek: that why i did say stop all running scripts on crossings
  • [2009/09/08 11:28] Teravus Ousley: therefore, there is no lag.. there's no serialization..
  • [2009/09/08 11:28] Simon Linden: right, that's the big win
  • [2009/09/08 11:28] Arawn Spitteler: The running scripts don't know they're crossing, until the crossing's been completed.
  • [2009/09/08 11:28] Teravus Ousley: there is no border
  • [2009/09/08 11:29] Simon Linden: I wonder how much stuff we'd break -- scripts that have the 256 size hard-coded in it
  • [2009/09/08 11:29] Teravus Ousley: it just looks 'visually' in the client, as if there is a border.
  • [2009/09/08 11:29] Teravus Ousley:
  • [2009/09/08 11:29] Susie Chaffe: in terms of vehicle scripts not many Is my guess Simon
  • [2009/09/08 11:30] Meeter: Timecheck : office hours is half over
  • [2009/09/08 11:30] Arawn Spitteler: I'm not sure scripts ahve the 256 limit hard coded in. If we don't get handed off, the numbers just continue.
  • [2009/09/08 11:30] xstorm Radek: and griefers can not use the crossing hole to make there push griefer prims work in places like korea1 for there will not be that crossing gap
  • [2009/09/08 11:30] Techwolf Lupindo: What is sounds like to me, everything was coded with 256 in mind. If those variables was just changed to int 32 or something, regines would be bounded by hardware, not boarders.
  • [2009/09/08 11:30] Simon Linden: There would be a design choice somewhere ... do you really spoof the viewer and have it think everything is still 256x256, or allow coordinates to get larger
  • [2009/09/08 11:31] xstorm Radek: there may be a problem with that
  • [2009/09/08 11:31] Simon Linden: I once did a search for 256 in our source code, and was not happy with how much it's baked in... :(
  • [2009/09/08 11:31] Techwolf Lupindo: If they limit is raised from 256, code everything to handle FFFF instead of FF.
  • [2009/09/08 11:31] Arawn Spitteler: If the pseudo-null-sim is passing the object to itself, then the viewer could be updated at liesure
  • [2009/09/08 11:32] xstorm Radek: yes 256 is a base thats been used to death
  • [2009/09/08 11:32] Arawn Spitteler: Getting an image of the server and system code, to map where the numbers luirk, is going to be the challenge
  • [2009/09/08 11:32] Teravus Ousley: Well, the trick really.. is the uses RegionHandle position + relative Position
  • [2009/09/08 11:32] Teravus Ousley: 'the viewer uses'
  • [2009/09/08 11:33] Teravus Ousley: the relative position can be > 256m
  • [2009/09/08 11:33] Simon Linden: I think this is really one of those projects where someone just needs to start experimenting and see what breaks
  • [2009/09/08 11:33] Andrew Linden: well perhaps now that OpenSim has the megaregion capability we'll eventually get around to doing it in the SL server. That's a pretty big project though. We certainly don't have the dev resources to do it now.
  • [2009/09/08 11:33] Aimee Trescothick: 542 references to the number 256 in the current viewer source, although many of them are comments
  • [2009/09/08 11:34] Aimee Trescothick: and many of them are things like texture dimensions
  • [2009/09/08 11:34] Teravus Ousley: Just like you can go infinitely high on the Z, you can go infinately far on the X and Y.. as you get higher, rounding errors will increase.
  • [2009/09/08 11:34] Simon Linden: One of the other problems with the 'spoofing' approach is the back end ... would our 512x512 mega-region look like 4 regions in our internal database, for example? We really don't want to be adding to our database load at all
  • [2009/09/08 11:34] Rex Cronon: just replace 256 with a constnt name, and then u can give any vaue to that costant:)
  • [2009/09/08 11:34] Teravus Ousley: Yes, it would be some grid overhead
  • [2009/09/08 11:35] Simon Linden: Well, you'd want that magic number to be flexible ... get the region size when you log into it, and have the viewer do the right thing
  • [2009/09/08 11:35] Teravus Ousley: To the grid structure, it would look like individual regions
  • [2009/09/08 11:35] Aimee Trescothick: that will break anything that relys on it packing into a single byte
  • [2009/09/08 11:35] xstorm Radek: what if the server sees one size and then tells the client another size ?
  • [2009/09/08 11:35] Arawn Spitteler: Please, having learned from the Open Space NNormality, don't forget to budget for the Janitor?
  • [2009/09/08 11:35] Andrew Linden: I dunno Simon. If you're talking about making a 512x512 region (as opposed to a megaregion that spoofs as 4X4 square of legacy 256x256 regions) then I can make a long list of things I know will break.
  • [2009/09/08 11:35] Twisted Laws: region mini-map data packs locations into bytes
  • [2009/09/08 11:36] Andrew Linden: however, building the spoofer megaregion... I can also come up with a long list of things that would break.
  • [2009/09/08 11:36] Aimee Trescothick: exactly
  • [2009/09/08 11:36] Teravus Ousley: Yes, Aimee, one thing that I found limited by that was the CoarseLocation Message. But.. for those things you send them out from the appropriate 'virtual' region.
  • [2009/09/08 11:36] Simon Linden: yep, many things will break. That's why it's been so difficult to get into it...
  • [2009/09/08 11:36] Aimee Trescothick nods
  • [2009/09/08 11:36] Andrew Linden: Most of them have to do with singleton classes and global variables.
  • [2009/09/08 11:36] Latif Khalifa: CoarseLocationUpdate message needs to be taken out back and shot :P
  • [2009/09/08 11:36] Latif Khalifa: for using 1 byte per coordinate
  • [2009/09/08 11:37] Ardy Lay: Hehe, yeah
  • [2009/09/08 11:37] Teravus Ousley: hehe
  • [2009/09/08 11:37] xstorm Radek: ok what is only islands what made bigger ?
  • [2009/09/08 11:37] Andrew Linden is trying to remember who wrote that CoarseLocationUpdate stuff originally.
  • [2009/09/08 11:37] Aimee Trescothick: there's another 1108 occurences of the number 255 in the source also
  • [2009/09/08 11:37] xstorm Radek: aaakkk
  • [2009/09/08 11:37] Aimee Trescothick: though lots of those are to do with colour and things
  • [2009/09/08 11:37] xstorm Radek: ok what if only islands are made bigger ?
  • [2009/09/08 11:37] Latif Khalifa: that's wht minimap shows you at hieght 0 if you fly are above 1024
  • [2009/09/08 11:37] Arawn Spitteler has to nag Michael Linden about the Fall Through in The Wall: SVC-22 should be ritualy mentioned, but seems a problem, that the object is placed in the simulation prior to the entire object being received.
  • [2009/09/08 11:37] Meeter: http://jira.secondlife.com/browse/SVC-22
  • [#SVC-22] Vehicles crossing region borders aren't always treated as vehicles and can get incorrectly returned if the destination parcel is no-entry or parcel-full
  • [2009/09/08 11:38] Techwolf Lupindo: The mini-map get confused after a certion high due to another FF number in the message.
  • [2009/09/08 11:38] Aimee Trescothick: yeah, height is a single byte * 4
  • [2009/09/08 11:38] Aimee Trescothick: so it has a 4m resolution and goes up to 1024
  • [2009/09/08 11:38] Latif Khalifa: i'd append another 3 bytes to teh message so it can be kept backwards compatible
  • [2009/09/08 11:38] Latif Khalifa: the same way agent uuid were added
  • [2009/09/08 11:39] Aimee Trescothick: this stuff only applies when it's sent by UDP though
  • [2009/09/08 11:39] xstorm Radek: there is more then one style of Vehicles !!!
  • [2009/09/08 11:39] Andrew Linden: Aimee, you're still talking about CoarseLocationUpdate?
  • [2009/09/08 11:39] Aimee Trescothick: there's provision in the source for coarselocation to be sent by TCP without these restrictions
  • [2009/09/08 11:39] Aimee Trescothick: yes
  • [2009/09/08 11:39] Latif Khalifa: that way old viewers will not be confused, and those 3 original bytes can be used as 8 least significat bits of the 16 bit xyz
  • [2009/09/08 11:39] Teravus Ousley: Yes, CoarseLocationUpdate message
  • [2009/09/08 11:39] xstorm Radek: some vehicles are attachments
  • [2009/09/08 11:40] Andrew Linden: hrm... the provision for TCP message? That is in the SL client?
  • [2009/09/08 11:40] Aimee Trescothick: yeah
  • [2009/09/08 11:40] Latif Khalifa: you don't wantCoarseLocationUpdate over tcp
  • [2009/09/08 11:40] Andrew Linden: yes, changing the message to TCP and to use some other format is the way to go
  • [2009/09/08 11:40] Arawn Spitteler's planning to submit Noah's Ark to the Department of Mutant Vehicles, using llSetPos, rather than Dynamic
  • [2009/09/08 11:41] Aimee Trescothick: "// the new TCP coarse location handler node
  • class CoarseLocationUpdate : public LLHTTPNode"
  • [2009/09/08 11:41] Andrew Linden: I doubt there is TCP handling on the server yet, but I may be wrong. Perhaps we're moving that stuff to a "capabilities" service.
  • [2009/09/08 11:41] Latif Khalifa: i really see no reasong to go tcp/caps on CoarseLocation, its perfect for current udp messaging
  • [2009/09/08 11:42] Latif Khalifa: nobody cares if they get lost/out of order
  • [2009/09/08 11:42] Teravus Ousley: Actually.. the X resolution is just 256m, the Y resolution is just 256m @1m, and the Z resolution is @4
  • [2009/09/08 11:42] Techwolf Lupindo: UDP is good for state info that expires quickly or is out of date quickly. Many games use UDP to send position updates.
  • [2009/09/08 11:43] Latif Khalifa: Teravus, agent UUIDs were appended to the messages
  • [2009/09/08 11:43] Latif Khalifa: you could append another 3 bytes per entry in the same way
  • [2009/09/08 11:43] Latif Khalifa: to have 8 more significant bits
  • [2009/09/08 11:43] Latif Khalifa: in a backward compatible way
  • [2009/09/08 11:44] Teravus Ousley: interesting.
  • [2009/09/08 11:44] Andrew Linden: Yes, I think it is possible to change the message format by appending to the end -- without breaking compatibility with older SL client versions
  • [2009/09/08 11:44] Latif Khalifa: so right now it has two blocks
  • [2009/09/08 11:44] Latif Khalifa: 1 positions
  • [2009/09/08 11:44] Latif Khalifa: 2 agent uuids
  • [2009/09/08 11:44] Latif Khalifa: you add 3 ) postitionEx (1 byte = 8 most significat bits) per coordinate
  • [2009/09/08 11:45] Latif Khalifa: so the viewer that knows about it will add 1) pos + 3) pos *256
  • [2009/09/08 11:45] Latif Khalifa: old viewer would just use 1) and be happy with it
  • [2009/09/08 11:46] Andrew Linden: right, that sounds doable
  • [2009/09/08 11:46] Ardy Lay: No more hiding bot-farms in the Z-axis.
  • [2009/09/08 11:47] Teravus Ousley: cool, that would save us from the following code also: http://opensimulator.org/viewgit/?a=commitdiff&p=opensim&h=855fb58c9650086122b76c586c4a4a408343143f#OpenSim%2FRegion%2FCoreModules%2FWorld%2FLand%2FRegionCombinerModule.cs
  • [2009/09/08 11:47] Andrew Linden: So, perhaps someone should create that proposal as a jira item, and add me as a watcher.
  • [2009/09/08 11:47] Andrew Linden: Or just assign it to me.
  • [2009/09/08 11:47] Teravus Ousley: (ugly code that sends the CoarseLocationUpdate through the proper 'virtual region'
  • [2009/09/08 11:47] Teravus Ousley: )
  • [2009/09/08 11:47] Aimee Trescothick thinks people should be allowed bot farms only on damage enabled land, they make fun cannon fodder
  • [2009/09/08 11:48] Andrew Linden: I'll see if we can get it done somehow.
  • [2009/09/08 11:48] Aimee Trescothick: LOL, last two lines make a great out of context quote :D
  • [2009/09/08 11:48] Arawn Spitteler: Oh, the solution of Bot Farms, is to charge people for Traffic; after all, builds are to be seen.
  • [2009/09/08 11:49] Ardy Lay: Many are social areas
  • [2009/09/08 11:49] Andrew Linden: I was under the impression that problems with bot farms had significantly decreased. Is that true?
  • [2009/09/08 11:49] Teravus Ousley: there are other issues.. like Parcels.. and Terrain.. but.. everything for the most part can be handled by getting the viewer to think that it's multiple regions
  • [2009/09/08 11:49] Ardy Lay: That would eat them alive.
  • [2009/09/08 11:50] Aimee Trescothick: I did push a small one into the sea a few hours ago
  • [2009/09/08 11:50] Aimee Trescothick: (bot farm that is)
  • [2009/09/08 11:50] Ardy Lay: Andrew, I don't go hunting for them but I did fall over something that MIGHT be a bot far but I could not determine what their altitude was so I marched on.
  • [2009/09/08 11:50] Arawn Spitteler imagines all those bots being scripted to AR
  • [2009/09/08 11:51] Aimee Trescothick: lol, please don't
  • [2009/09/08 11:51] Teravus Ousley: err, yeah.. that would be bad
  • [2009/09/08 11:51] Rex Cronon: imagine all of them trying to IM u:)
  • [2009/09/08 11:51] Latif Khalifa: updating coarse location would allow fixing some long standing viewer bugs irrespective of mega region issue
  • [2009/09/08 11:52] Latif Khalifa: i'll write up a jira proposal
  • [2009/09/08 11:52] Arawn Spitteler doen't think malware bots to be unprecedented
  • [2009/09/08 11:52] Teravus Ousley: cool ;)
  • [2009/09/08 11:52] Simon Linden: yeah, I think we really need to come up with a replacement message that has a better design and is extensible, hopefully phasing out the old one eventually
  • [2009/09/08 11:53] Arawn Spitteler: Replacement Message?
  • [2009/09/08 11:53] Simon Linden: One message to rule them all :)
  • [2009/09/08 11:53] Andrew Linden: ImprovedCoarseUpdate
  • [2009/09/08 11:53] Aimee Trescothick: well, the code is there for the HTTP one, so presumably for better or worse that decision has already been taken?
  • [2009/09/08 11:54] Aimee Trescothick: or maybe taken and then forgotten :)
  • [2009/09/08 11:54] Andrew Linden: Yes, I wonder.
  • [2009/09/08 11:54] Andrew Linden: There is so much development going on now it is hard to stay informed about all of it.
  • [2009/09/08 11:54] Arawn Spitteler: Taken and Shelved? We're still waiting on SVC-93 where someone borked the rotation code, and couldn't sort out where.
  • [2009/09/08 11:55] Meeter: http://jira.secondlife.com/browse/SVC-93
  • [#SVC-93] llSetPrimitiveParams PRIM_ROTATION and llSetRot incorrectly implemented for child prims
  • [2009/09/08 11:55] Meeter: Timecheck : office hours is almost over
  • [2009/09/08 11:55] Arawn Spitteler: Then, we should also mention svc-2931 where arbitrary decisions are not user friendly.
  • [2009/09/08 11:55] Meeter: http://jira.secondlife.com/browse/SVC-2931
  • [#SVC-2931] llSetLinkPrimitiveParams content breakage, only moves agents 54 meters now
  • [2009/09/08 11:56] Andrew Linden: Thanks for all those reminders Arawn.
  • [2009/09/08 11:56] Aimee Trescothick: it may have been done just to send the UUIDs for friend identification on the minimap, as originally they didn't think it would go in the UDP version ... until I pointed out they were actually being sent and it was working just fine lol
  • [2009/09/08 11:56] Arawn Spitteler: I'm wondering, why vehicles, crossing boundaries, don't get checked for numbers of links, before rezzing?
  • [2009/09/08 11:56] Andrew Linden: I'm going to have to leave right at 12:00. I've got another meeting starting then.
  • [2009/09/08 11:57] Rex Cronon: i hopped u could mov at lest 256::(
  • [2009/09/08 11:57] Liisa Runo: and 4096 up
  • [2009/09/08 11:57] Arawn Spitteler wonders how to qualify as a junior developer
  • [2009/09/08 11:58] Arawn Spitteler: Above 4096, we've been known to go, but I don't recall arrows going upside down back then.
  • [2009/09/08 11:59] Arawn Spitteler: We didn't have arrows, back then
  • [2009/09/08 11:59] Andrew Linden: A junior developer would typically be a Computer Science student just graduating, or a developer with less than 5 years work experience.
  • [2009/09/08 12:00] Meeter: Thank you for coming to Linden office hours
  • [2009/09/08 12:00] Arawn Spitteler has math degree, and experience in idle philosophy
  • [2009/09/08 12:00] Arawn Spitteler: I'll have to se what it takes to graduate in Computer Science, and demonstrate some initiativ e
  • [2009/09/08 12:01] Aimee Trescothick: lack of sleep and a good sense of humour
  • [2009/09/08 12:01] Andrew Linden: I didn't know you had a math degree Arawn. Cool.
  • [2009/09/08 12:01] Teravus Ousley: hehe (sense of humor)
  • [2009/09/08 12:01] Arawn Spitteler: Oh, 12:01, time to kick Andrew out, and mob Simon. NU'89
  • [2009/09/08 12:01] Andrew Linden: Yes, I've got to go.
  • [2009/09/08 12:01] Simon Linden: I gotta RUN...
  • [2009/09/08 12:01] Andrew Linden: thanks for coming everyone
  • [2009/09/08 12:01] Rex Cronon: tc andrew
  • [2009/09/08 12:02] Aimee Trescothick: bye :)
  • [2009/09/08 12:02] Simon Linden: Thanks all for coming, see you next time