AW Groupies/Chat Logs/AWGroupies-2009-02-17

• [2009/02/17 9:30] Morgaine Dinova: Hi Latha
• [2009/02/17 9:30] Morgaine Dinova: Hi Latha
• [2009/02/17 9:31] Morgaine Dinova: Hiya Dahlia
• [2009/02/17 9:31] Dahlia Trimble: Hi :)
• [2009/02/17 9:32] Aleric Inglewood: accepted your inventory offer.
• [2009/02/17 9:33] Morgaine Dinova: Not gingerbread men. Anything but evil gingerbread men!!!
• [2009/02/17 9:33] Morgaine Dinova: Woot! Go get'em Latha!
• [2009/02/17 9:35] Aleric Inglewood: Okay.. .first proposal: SOMETHING to synchronize gestures with sounds! (and not play sounds 10 seconds too late) .. (ok, just joking, I know it's now about that kind of "proposals" now ;)
• [2009/02/17 9:35] Aleric Inglewood: isn't not*
• [2009/02/17 9:35] Zha Ewry: OK, I think we can call time and assume we have the bulk of people who are going to show.
• [2009/02/17 9:35] Jonit Ivory: greetings all
• [2009/02/17 9:35] Zha Ewry: So.. Greeetings one and all
• [2009/02/17 9:35] Morgaine Dinova: Hehe
• [2009/02/17 9:36] Morgaine Dinova: Greetings Zha :-)
• [2009/02/17 9:36] Dahlia Trimble: waves
• [2009/02/17 9:36] Zha Ewry: For those whoe are new, and even those are are not...
• [2009/02/17 9:36] Junta Kohime: accepted your inventory offer.
• [2009/02/17 9:37] Zha Ewry: This is the ongoing discussion of things interioperable and virtual world architecturish parallel to Linden's office hour structure
• [2009/02/17 9:37] Zha Ewry: Since the center of mass of that seems to be headed to the IETF, this is going to largely focus on IETF realted stuff in the near term
• [2009/02/17 9:38] Morgaine Dinova: Fair enough. With the occasional local stuff too, if important, hopefully. Or if fun ;-)))
• [2009/02/17 9:38] Zha Ewry: Anyone who hasn't looked that MMOX BOF at IETF 27(?) in San Francisco ought to look it up. We're now officially calendared
• [2009/02/17 9:39] Saijanai Kuhn: yay
• [2009/02/17 9:39] Jonit Ivory: very good :)
• [2009/02/17 9:39] Morgaine Dinova: purrs at Imaze :-)
• [2009/02/17 9:40] Zha Ewry: Given that, I wanted to biefly talk about the IETF and the goals of an IETF BOF, and what we might expect to get out of all of this
• [2009/02/17 9:40] Zha Ewry: I need to preface this with the usual "I don't speak for Linden, I barely spek form my own employers, and on bad days, I'm nore sure I speak coherently at all ;-) ) "
• [2009/02/17 9:41] Jonit Ivory: always with coherence
• [2009/02/17 9:41] Zha Ewry: The IETF, is the Internet Engineering Task Force
• [2009/02/17 9:41] Morgaine Dinova: Zha, I made a quickie page to help out documentation control overnight --- most of the current useful MMOX links are there -- https://wiki.secondlife.com/wiki/MMOX
• [2009/02/17 9:41] Zha Ewry: basically, the folks who try to manage the low level and mid-level stuff which forms the backbone of the internet
• [2009/02/17 9:41] Rex Cronon: hi everyboy
• [2009/02/17 9:42] Zha Ewry: Minor stuff like Domain Name Service, TCP/IP the RFC822 spec for e-mail, etc.
• [2009/02/17 9:42] Dahlia Trimble: hiya Rex
• [2009/02/17 9:42] Morgaine Dinova: Hiya Rex
• [2009/02/17 9:42] Rex Cronon: hiii
• [2009/02/17 9:42] Zha Ewry: So.. the reason the IETF makes a lot of sense is that what we need, is some base building blocks we all can share. The good news about goign to the IETF is they're goof at that, and it gives us a formal
• [2009/02/17 9:43] Zha Ewry: not run by any single company place to work on stuff
• [2009/02/17 9:43] Dahlia Trimble: Saijanai is in spam mode today
• [2009/02/17 9:43] Zha Ewry: The bad news, is that it is probably slightly less directly coordinated with the day to day work at Linden and possible OpenSim as well
• [2009/02/17 9:44] Latha Serevi: The spam quantity is hardly up to Sai's high standard yet.
• [2009/02/17 9:44] Aleric Inglewood: The name IETF is not a garantee this will work out at all. I've been in an IETF BOF before (to formalize the next generation IRC protocol) and that failed horrible.
• [2009/02/17 9:44] Zha Ewry: Right
• [2009/02/17 9:44] Anthonys Republic: accepted your inventory offer.
• [2009/02/17 9:44] Zha Ewry: So.. what a BOF is about is trying to see if there is in fact:
• [2009/02/17 9:44] Zha Ewry: A community which wants to work on stuff
• [2009/02/17 9:44] Zha Ewry: A coherent idea of what it woudl work on
• [2009/02/17 9:44] Zha Ewry: and enough overla with what the IETF does, for it to make sense
• [2009/02/17 9:44] Dahlia Trimble: how did it fail? was the standard adopted?
• [2009/02/17 9:45] Jonit Ivory: do you see any RFCs that Linden get out of involving teh IETF as pecursors to openin sourcing teh server code?
• [2009/02/17 9:45] Rex Cronon: come on. don't pick on sai for trying to get more people interested in mmox:)
• [2009/02/17 9:45] Jonit Ivory: open^^
• [2009/02/17 9:45] Latha Serevi: Rex, that was praise. Couldn't you tell?
• [2009/02/17 9:45] Zha Ewry: I see it as othogonal, and if you want to know Linden Lab's plans for such things, they are the right people to talk to, not me. ;-(
• [2009/02/17 9:45] Jonit Ivory: indeed
• [2009/02/17 9:46] Zha Ewry: In general, the two topics cross cut each other
• [2009/02/17 9:46] Aleric Inglewood: Dahlia: everyone involved wanted something else and couldn't agree... there was no structure in the organization and when it became clear it was just all going to be a waste of time, the smartest people left first and the whole thing bled to death (after a few years even the last website traces disappeared).
• [2009/02/17 9:46] Zha Ewry: Standards tell you how to interconnect, in absence of a single code base
• [2009/02/17 9:46] Morgaine Dinova: Ouch Aleric
• [2009/02/17 9:47] Zha Ewry: and open code base, lets you share coding effort.
• [2009/02/17 9:47] Latha Serevi: I'm interested in delimiting the scope of the IETF stuff -- particularly, whether the 3-d and scripting and other high-level stuff might be usefully excluded and sent to MPEG-V or a non-standards-body approach. Or at least if this year's focus can be narrowed a bit.
• [2009/02/17 9:48] Latha Serevi: Is IETF for authentication and plumbing?
• [2009/02/17 9:48] Morgaine Dinova: Latha: Infinity said to us last night thast she would answer our questions about scope.
• [2009/02/17 9:48] Zha Ewry: I tend tot hink of the IETF as the plumbing and interop and wire format space
• [2009/02/17 9:48] Latha Serevi: Yes, but I think it was wrong-headed to expect Infinity to lecture us -- WE decide
• [2009/02/17 9:48] Dahlia Trimble: so LL will define scope?
• [2009/02/17 9:48] Zha Ewry: I wouldn't expect to see things that primarily happen on the client, or inside the server to be of as much intereest
• [2009/02/17 9:48] Zha Ewry: NO
• [2009/02/17 9:48] Zha Ewry: The IETF charter will
• [2009/02/17 9:49] Zha Ewry: and that's a jointly developed work product of the people who assemble into the BOF and the mailing list and the working group
• [2009/02/17 9:49] Dahlia Trimble: forgot to duck... :/
• [2009/02/17 9:49] Zha Ewry: Linden, gets a say, roughtly proportional tot he amoutn of effort they put into that
• [2009/02/17 9:49] Jonit Ivory: and they have the resources to steamroller it?
• [2009/02/17 9:49] Morgaine Dinova: Latha: correct, but LL's proposal is the only one on the table atm --- so I wanted (because Jon Watte asked specifically) the OGP designers to explain what was covered and what not, in THEIR proposal (only).
• [2009/02/17 9:50] GG Office: Hours HUD v1.5: 10:00 AM: Lexie Linden: SL Volunteering (Mentors Only)
• http://slurl.com/secondlife/SLVEC/190/144/22 (Starts in 10 minutes)
• [2009/02/17 9:50] GG Office: Hours HUD v1.5: 10:00 AM: Lexie Linden: SL Volunteering (Alternative Location)
• http://slurl.com/secondlife/Ambleside/181/99/30 (Starts in 10 minutes)
• [2009/02/17 9:50] Morgaine Dinova: Doesn't mean we'll stick to their proposal;.
• [2009/02/17 9:50] Zha Ewry: Well, thus far, Linden's been about 1/4 of infinity's time, and about 10% of Zero's time
• [2009/02/17 9:50] Zha Ewry: I don't feel steamrolled
• [2009/02/17 9:50] Jonit Ivory: it is in that case a natural progression to obtain standards, salutable
• [2009/02/17 9:51] Zha Ewry: and. for those wondering fresh from my cell phone...
• [2009/02/17 9:51] Morgaine Dinova: Ditto, I don't feel steamrolled (but I do feel Lindens are not putting in much presence). However, I most importantly want to see an independent proposal from Opensim.
• [2009/02/17 9:51] Zha Ewry: Infinity is delayed with a RL wooshing teelport bar
• [2009/02/17 9:51] Aleric Inglewood: Are decisions going to be made by means of voting? If not, how are they going to be made, and if so, who gets to vote?
• [2009/02/17 9:51] Zha Ewry: IE, traffic jam
• [2009/02/17 9:51] Zha Ewry: The IETF works by "rough consensus"
• [2009/02/17 9:52] Morgaine Dinova: I think we're trying to arrive at a technical consensus. Voting is only required in the event of a dispute.
• [2009/02/17 9:52] Morgaine Dinova: Agreed with Zha.
• [2009/02/17 9:52] Zha Ewry: There is a pile of IETF process documentation,if we getforced into that, we've failed
• [2009/02/17 9:52] Rex Cronon: voting is also a way to find out what others are thinking:)
• [2009/02/17 9:52] Zha Ewry: As Morgaine has noticed
• [2009/02/17 9:53] Latha Serevi: Zha, how are you personally seeing this process going so far, and what goals do you have for this month's pre-meeting discussion?
• [2009/02/17 9:53] Zha Ewry: we have a couple of peopel from a very NON linden perpsectiveengaged
• [2009/02/17 9:53] Junta Kohime: dinner time for me, bye
• [2009/02/17 9:53] Aleric Inglewood: I agree too that consensus is better - but sometimes an official vote can help as milestone to make progress.
• [2009/02/17 9:53] Zha Ewry: Jon Watte, the CTO of Forterra for example
• [2009/02/17 9:53] Dahlia Trimble: bye Junta :)
• [2009/02/17 9:53] Rex Cronon: tc
• [2009/02/17 9:53] Imaze Rhiano: bye Junta
• [2009/02/17 9:53] Zha Ewry: And, yeah, the reason standard bodies sometimes vote, is that votes flush out formal positiones
• [2009/02/17 9:54] Zha Ewry: A lto of IETF groups tend so hold straw polls for that purpose, and I'd expect we woudl do that too
• [2009/02/17 9:54] Latha Serevi: Aside from collecting people into a room, don't we want to be a bit more focused about some candidate 2009 places to put our standards efforts?
• [2009/02/17 9:54] Zha Ewry: So.. What I'm looking for fromt his is three things
• [2009/02/17 9:55] GG Office: Hours HUD v1.5: 10:00 AM: Lexie Linden: SL Volunteering (Mentors Only)
• http://slurl.com/secondlife/SLVEC/190/144/22 (Starts in 5 minutes)
• [2009/02/17 9:55] GG Office: Hours HUD v1.5: 10:00 AM: Lexie Linden: SL Volunteering (Alternative Location)
• http://slurl.com/secondlife/Ambleside/181/99/30 (Starts in 5 minutes)
• [2009/02/17 9:55] Zha Ewry: Getting people from outside the insualr Linden/OpenSim community to come and play seriously, starting from the current assumptions and saying "This doesn't work for us, because we need to be able to do X"
• [2009/02/17 9:56] Zha Ewry: Getting some very basic low level building blocks on the path to a formal spec
• [2009/02/17 9:56] Zha Ewry: Second generatinos of LLSD, Constructive Geometry (With a proper extensions model) and such
• [2009/02/17 9:56] Zha Ewry: and agreeing to a general set of terms and servcies which comprise a virtual world
• [2009/02/17 9:58] Zha Ewry: So, for example, 90% of the virtual worlds do some form of a stream of "X has changed" updates between components. But, we don't in general share a common way of talkign about that
• [2009/02/17 9:59] Zha Ewry: Closer to home...
• [2009/02/17 9:59] Zha Ewry: I also expect to get things like "Hypergrid" vs OGP
• [2009/02/17 9:59] Zha Ewry: on the table
• [2009/02/17 9:59] Dahlia Trimble: are there examples of standards which are successful but defined before wide adoption or implementation?
• [2009/02/17 10:00] GG Office: Hours HUD v1.5: 10:00 AM: Lexie Linden: SL Volunteering (Mentors Only)
• http://slurl.com/secondlife/SLVEC/190/144/22 (Starts now)
• [2009/02/17 10:00] GG Office: Hours HUD v1.5: 10:00 AM: Lexie Linden: SL Volunteering (Alternative Location)
• http://slurl.com/secondlife/Ambleside/181/99/30 (Starts now)
• [2009/02/17 10:00] Zha Ewry: and get some of the building blocks for naming inventory and assets which cross grids dsicussed, along with the naming and policy stuff that ties to that
• [2009/02/17 10:00] Zha Ewry: chuckles softly
• [2009/02/17 10:00] Zha Ewry: NFS comes to mind
• [2009/02/17 10:00] Dahlia Trimble: I thought nfs was implemented first
• [2009/02/17 10:00] Zha Ewry: gestures at the sim we're on
• [2009/02/17 10:00] Zha Ewry: At about the level the SL is
• [2009/02/17 10:01] Zha Ewry: NFS was pretty close to a single platform solutino until the PC and non *NIX folks came to play
• [2009/02/17 10:01] Aleric Inglewood: Would git be useful (asking because git represents a "file system" that can only be added to, not changed). The advantage is, of course, that lag and meshed linked are never a problem when retrieving content.
• [2009/02/17 10:02] Aleric Inglewood: meshed links*
• [2009/02/17 10:03] Aleric Inglewood: Other advantages is that it is wellknown, robust and distributed (scalable)
• [2009/02/17 10:03] Zha Ewry: There are, dependingo n how you count them, dozens of virtual worlds, and if you look at ones which are "close" to SL
• [2009/02/17 10:03] Zha Ewry: at least a half doezen which would form a good critcal mass for a shared space
• [2009/02/17 10:03] Zha Ewry: ie. if one can get shared bits between
• [2009/02/17 10:04] Zha Ewry: Second Life, OpenSim, Forterra, and things liek Croquet, you actually start to have a generalized set of buildnig blocks
• [2009/02/17 10:04] Jonit Ivory: you seem positive about teh IETF move by LL, any glaring reservations?
• [2009/02/17 10:04] Rex Cronon: just a sec. sl and os ARE the same
• [2009/02/17 10:05] GG Office: Hours HUD v1.5: 10:00 AM: Lexie Linden: SL Volunteering (Mentors Only)
• http://slurl.com/secondlife/SLVEC/190/144/22 (Started 5 minutes ago)
• [2009/02/17 10:05] GG Office: Hours HUD v1.5: 10:00 AM: Lexie Linden: SL Volunteering (Alternative Location)
• http://slurl.com/secondlife/Ambleside/181/99/30 (Started 5 minutes ago)
• [2009/02/17 10:05] Morgaine Dinova: In answer to the implicit "Will anyone follow what we suggest?" ... I think we have to recognize that this is just round #1, and it may involve only those worlds that help out. The future is uncertain ... or I'd be betting money ;-)
• [2009/02/17 10:05] Zha Ewry: Reservations?
• [2009/02/17 10:05] Jonit Ivory: reasons why you are not comfortable with it
• [2009/02/17 10:05] Zha Ewry: Standards processes are slow, painful, and go in unexpected directions
• [2009/02/17 10:05] Creative Acronym: accepted your inventory offer.
• [2009/02/17 10:06] Dahlia Trimble: not really Rex, Opensim is attempting to be a superset of sl
• [2009/02/17 10:06] Object: Touched.:
• [2009/02/17 10:06] [[User:Object: <-0.94339,|Object: <-0.94339,]]: 0.14065, 0.30039>
• [2009/02/17 10:06] Zha Ewry: If you count the RFCS, the IETF has done, delete the duplicates, which are follow ons, and then count up how many are used, it's a slightly depressing ratio
• [2009/02/17 10:06] Rex Cronon: os, uses the same viewer, therefore has to use the same protocols
• [2009/02/17 10:07] Zha Ewry: That said, the IETF has a better sucess rate than a lot of groups
• [2009/02/17 10:07] Hermit Barber: Or offer a bride ort gateway.
• [2009/02/17 10:07] Dahlia Trimble: opensim has hooks for alternative protocols
• [2009/02/17 10:07] Zha Ewry: nods at Dahlia
• [2009/02/17 10:07] Morgaine Dinova: Aleric: VWs suffer from an immense data explosion as they stand, even with object mutation. If objects were immutable like in Git, I think we'd have meltdown. It's not helpful in any case. Only the Wayback Machine wants old world state ;-)
• [2009/02/17 10:07] Rex Cronon: for all we know ll servers could have similar hooks too:)
• [2009/02/17 10:07] Zha Ewry: And, increasingly people are willing to support paired changes to the client and the OpenSim base, which allows it to divergefurther
• [2009/02/17 10:08] Dahlia Trimble: well we know it with opensim ;)
• [2009/02/17 10:08] Zha Ewry: My personal take (which influences, but is not my employers take)
• [2009/02/17 10:08] Zha Ewry: is that we will push on the IETF work, in parallel with OpenSim work, and we will see what happens
• [2009/02/17 10:09] Zha Ewry: Linden has said they will do this as "good faith"
• [2009/02/17 10:09] Zha Ewry: Some part of the OpenSim community will likely try and stay in synch
• [2009/02/17 10:09] Aleric Inglewood: Morgaine: Ok. Nevertheless, I'm slightly concerned with non-distributed repositories. If the asset server is stuck on a simple "database model" it will break down... heh, actually, we all know it is ALREADY breaking down. I also look at the problems they had/have with sourceforge and CVS / SVN.
• [2009/02/17 10:10] Zha Ewry: If we actually get players like Forterra to engage in the exercise
• [2009/02/17 10:10] Zha Ewry: I think it becomes compelling
• [2009/02/17 10:10] Morgaine Dinova: I further propose that we also work on independent implementations, purely from our MMOX specs, both to prove the point that the documents we make can stand alone, and also to satisfy the IETF requirement for reference implementations.
• [2009/02/17 10:10] FWord Utorid: whatever happened to the one world grid where we could tp back and forth? or is that now in the red tape menagerie?
• [2009/02/17 10:11] Morgaine Dinova: That's the goal of MMOX, FWord.
• [2009/02/17 10:11] FWord Utorid: ok. red tape menagerie. rename, stall, take it to committee. go!
• [2009/02/17 10:11] Zha Ewry: And, its largely dependent on people like Linden feeling there are specs which let them do that without surrenduring control of assets and thinsg they care about
• [2009/02/17 10:11] Goldie Katsu: I like the independent implementaitons concept
• [2009/02/17 10:12] FWord Utorid: goldie: yup
• [2009/02/17 10:12] Zha Ewry: About 50% of the problems are technical, about 50% are much more policy, politics and legal
• [2009/02/17 10:12] Goldie Katsu: Ideally there will be ways to interop test once the standards become more solidified.
• [2009/02/17 10:12] Ellla McMahon: sorry, gth .. thank you everyone :)
• [2009/02/17 10:12] Rex Cronon: tc
• [2009/02/17 10:12] Morgaine Dinova: waves to Ellla
• [2009/02/17 10:12] Zha Ewry: The IETF, is strongly driven by "bake off" style testing
• [2009/02/17 10:13] FWord Utorid: zha: I would say the pie chart is skewed more towards the political, but we're all enjoying the show.
• [2009/02/17 10:13] Zha Ewry: No RFC is going to be ratfied without some deomnstration of workign interoperable code
• [2009/02/17 10:13] FWord Utorid: idk, an rfc for this stuff would sorta kinda be insane
• [2009/02/17 10:13] Zha Ewry: why?
• [2009/02/17 10:14] Morgaine Dinova: I don't see much politics yet. And especially none from LL --- they're being very all-embracing so far, I'm impressed. Time will tell of course.
• [2009/02/17 10:14] Aleric Inglewood: Zha: I've spoken with a developer of OpenSpace (I think it was) and they want to try a model without currency or economy. That means that everything is Free (as in GPL). THAT would result to a TOTAL different way of storing and distributing assets imho - how can we ever support both with one protocol?
• [2009/02/17 10:14] Zha Ewry: Nope
• [2009/02/17 10:14] Zha Ewry: That;'s actually not at all true
• [2009/02/17 10:14] FWord Utorid: zha, idk, the old school rfcs you usually look at are for really basic protocols. this is sort of a slam everything together into one basket scheme and it could take a while to write it all down
• [2009/02/17 10:14] Zha Ewry: if you seperate mechanism from the policy
• [2009/02/17 10:15] Zha Ewry: you move the asset idetnically either way
• [2009/02/17 10:15] Jonit Ivory: I think an rfc is "exactly" what it needs, technically at least
• [2009/02/17 10:15] Zha Ewry: in one case you need to include some tokens to prove that you are permitted to get the asset
• [2009/02/17 10:15] Zha Ewry: in the other you don't.
• [2009/02/17 10:15] Zha Ewry: But, if we have an RFC that says
• [2009/02/17 10:15] Zha Ewry: "Here is how you move the bits"
• [2009/02/17 10:15] Zha Ewry: we seperate out those issues form each other nicely
• [2009/02/17 10:15] Zha Ewry: if, on the other hand we munge it all tgether
• [2009/02/17 10:16] Zha Ewry: then there is never a chance of letting the two worlds share at all
• [2009/02/17 10:16] FWord Utorid: sure, not objecting to the rfc, just saying it will be a pretty different one.
• [2009/02/17 10:16] Zha Ewry: Not one RFC, a set of them
• [2009/02/17 10:16] Aleric Inglewood: I disagree - if everthing is free, you could store assets distributed on the clients and retrieve information in some fuzzy way (depending on the large numbers of people connected), while an economy dictates central control of some asset server - it must have a major influence on the protocol imho.
• [2009/02/17 10:16] Morgaine Dinova: Aleric: I think there's no difference at all, Aleric. SL already has totally unencumbered assets within its own grid, and it takes very little to add an "Free Interop Accept" bit to allow creator's items to move freely to the worlds you describe.
• [2009/02/17 10:16] Zha Ewry: so you use the parts which you overlasp on, and differe when you differ
• [2009/02/17 10:16] Rex Cronon: so u just want to provide protocols for trasport only, and none for security?
• [2009/02/17 10:16] Zha Ewry: brb, have to cop with a RL distrction
• [2009/02/17 10:18] Morgaine Dinova: Rex: security is certainly being handled in the protocols. Of course it's not the security of single-provider walled gardens, but it will express the intentions of creators and owners to whatever extent is possible.
• [2009/02/17 10:19] Latha Serevi: As I see it, VW interop needs about half a dozen RFC's, for the plumbing, plus some higher-level stuff to be developed by other processes. MMOX is the group for scoping out the half dozen, and hopefully to help aim a few people towards the non-IETF bits.
• [2009/02/17 10:19] FWord Utorid: yup
• [2009/02/17 10:19] Dahlia Trimble: I thought the goal of this effort was to be able to handle a spectrum of security implementations
• [2009/02/17 10:19] Goldie Katsu: I think different models for assets shoudl be supported, as well as mechanisms for permission marking (and for different domains to share info that helps determine trust/what they are likely to share.)
• [2009/02/17 10:19] BlueWall Slade: OpenSim Interop, Gateway 7000 Fri Feb 20 18:00 ‚ Fri Feb 20 19:00 secondlife://ucigrid04.nacs.uci.edu:9007/
• [2009/02/17 10:20] Morgaine Dinova: As long as people understand that there are no magic bullets possible for asset protection once it goes beyond a single managed domain, we'll be fine. Of course if people are not well informed about what to expect, then there could be a problem. So we have to be clear.
• [2009/02/17 10:20] FWord Utorid: i don't see one meatball big enough to maneuver everything into teh square peg round hole conundrum
• [2009/02/17 10:20] BlueWall Slade: Interop meeting on OSGrid/Hypergrid to UCI Gateway 7000
• [2009/02/17 10:21] Morgaine Dinova: Goldie: yep
• [2009/02/17 10:21] Latha Serevi: Regarding what kinds of stuff to do inside the IETF process, I liked the sound of Zha's "X has changed" updates between components; but I would think that Constructive Geometry is out of scope. Perhaps X can be left undefined. Comments?
• [2009/02/17 10:21] Zha Ewry: Back
• [2009/02/17 10:21] Morgaine Dinova: Interop meeting now, on Opensim? That's a bit silly isn't it, to plan a clash?
• [2009/02/17 10:21] Goldie Katsu: Yeah the only magic bullet is not to interconnect outside of your locally controlled domain. But even that can benefit from interoperability standards.
• [2009/02/17 10:22] Hermit Barber: Intellectual property controls should be implemented as a security object embedding other objects and exposing some set of the attributes. If an object requestor is able to unwrap the object, it can use it. If not it cannot. The evaluation can be made before or after transfering the object. This ought to prove no obstacle to interoperability at all as at the moment you can request things in your inventory and the asset server says "Geh?" and that's it. Yet we can still continue without that thing.
• [2009/02/17 10:22] Zha Ewry: I'd argue, at some length, and tedious enthusiasm, that the bulk of interop is orthoganal to the security discussion
• [2009/02/17 10:22] Shamir Katsu: There are models that can be used for domains that wish to protect themselves more, a transition zone, like what was done for classifying and declassifying on mandatory security systems.
• [2009/02/17 10:22] Rex Cronon: maybe that object should be given out only to specific users only
• [2009/02/17 10:22] Dahlia Trimble: there is the normal office hour at 11 on Wright Plaza, but I dont think the agenda has been decided
• [2009/02/17 10:23] Morgaine Dinova: Latha: I'm going to be fighting hard for extensibility, so my comment is: Yup ;-)
• [2009/02/17 10:23] Hermit Barber: If objects are active and extensible than you can leave it up to the objects as to what they will allow/support.
• [2009/02/17 10:23] Zha Ewry: I'd also argue, wit far less certainty that there are useful models which depend on trust and enforcement mechansism, and recognize the technical limitations
• [2009/02/17 10:23] Zha Ewry: but still support knowing what the creator's intent was, and defer to that
• [2009/02/17 10:24] Aleric Inglewood: Maybe I miss something... but I still think that a protocol can and will be highly depending on what politics demand. As an example, I once designed a protocol for authentication of clients with the demands: nobody can be trusted (not even the servers). <started to explain the result here, but it as too long>... it *certainly* was something that needed a ground-up design with the political demand as starting point. :/
• [2009/02/17 10:24] Zha Ewry: I'd look at the web, Aleric, we share HTTP, HTML, and all the underpinnings in both
• [2009/02/17 10:24] Rex Cronon: even in rl there are technical limitation. but that doesn't stop people from puting locks on their doors:)
• [2009/02/17 10:24] Zha Ewry: secured, isnside the firewall environements
• [2009/02/17 10:24] Zha Ewry: an in fully public spaces
• [2009/02/17 10:24] Hermit Barber: Zha: Why I like an executable security policy. Let the creator decide. As long as the policy is in a library that says what it does to humans and implements the same in code, where is the challenge?
• [2009/02/17 10:25] Zha Ewry: we eveb stream the same content over both HTTP and HTTPS contexts and render it identically on the client
• [2009/02/17 10:25] Morgaine Dinova: Aleric: I agree --- but we don't build the policies into the protocols, but merely provide the mechanism for implementing them. A provider can choose to reply Yes or No --- we just provide the means to ask the question and handle the transport.
• [2009/02/17 10:25] Zha Ewry: Its quite possible to seperate out the layers of the stack from each other, and not entangle the base represenstaion with the security layers
• [2009/02/17 10:25] Rex Cronon: that is expected zha:)
• [2009/02/17 10:25] Infinity Linden: it might be interesting to note, the security work item was defined as a separate document from the base standards. this allows us to define a security policy similar to what is deployed today, then develop a different or better one in the future
• [2009/02/17 10:26] Zha Ewry: looks up and wonders when infinite trouble arrived
• [2009/02/17 10:26] Goldie Katsu: chuckles
• [2009/02/17 10:26] Zha Ewry: Didn't see you come in Infinity :-)
• [2009/02/17 10:26] Aleric Inglewood: Morgain: That will only work for p2p protocol - which can in its very nature cripple the working of the whole.
• [2009/02/17 10:26] Infinity Linden: i sneak about from time to time, chasing my tail
• [2009/02/17 10:26] Goldie Katsu: looks for her tail
• [2009/02/17 10:26] Morgaine Dinova: Infinity: I made a quickie page to help out documentation control overnight --- most of the current useful MMOX links are there -- https://wiki.secondlife.com/wiki/MMOX
• [2009/02/17 10:26] Latha Serevi: There'll be some interesting discussion and decisions to make regarding whether security mechanisms come via capabilities you pass, or wrappers you unwrap, etc ... but I doubt we can get into it today! Pop back up to the top level, perhaps?
• [2009/02/17 10:27] Zha Ewry: Right, in fact, the prupose of the BOF is to lay out those discussions, nto have them
• [2009/02/17 10:27] Zha Ewry: "We need to understand how we're going to look at topic X"
• [2009/02/17 10:27] Zha Ewry: not solve topic X
• [2009/02/17 10:27] Zha Ewry: so.. an example I have been using
• [2009/02/17 10:27] Infinity Linden: and yeah... one of my favorite guiding principles in protocol design is... "define mechanism, not policy"
• [2009/02/17 10:27] Morgaine Dinova: Yep, to examine the problem space, not supply the solutions.
• [2009/02/17 10:27] Goldie Katsu: watches the gre gradually resolve with points of red.
• [2009/02/17 10:28] Zha Ewry: "We need to justify why LLSD is a good starting point" not "Lets argue over the management of 64 bit ints in LLSD"
• [2009/02/17 10:28] Morgaine Dinova: Yup
• [2009/02/17 10:28] Zha Ewry: And, yes, in general
• [2009/02/17 10:28] Zha Ewry: We're going to be trying very very hard to speerate out
• [2009/02/17 10:28] Infinity Linden: but that being said, we have three genetically diverse implementations of a protocol we would like to standardize. i would argue that we have a very good understanding of aspects of virtual worlds sufficient to warrant a base protocol
• [2009/02/17 10:28] Rex Cronon: how can u only define mechanism and ignore the policy?
• [2009/02/17 10:28] Dahlia Trimble: <-- afk but will read the scrollback upon return
• [2009/02/17 10:29] Zha Ewry: mechanism from policy from implementation
• [2009/02/17 10:29] Infinity Linden: i thought it was 128 bit ints in LLSD
• [2009/02/17 10:29] Infinity Linden: @Rex... look at X/Windows
• [2009/02/17 10:29] Zha Ewry: So, put in slots in the mechanism for people to include autheticatiohn tokens, seperate from defining them
• [2009/02/17 10:29] Rex Cronon: which part infinity:)
• [2009/02/17 10:29] Zha Ewry: Put in places for where the policy expressions go, seperate from the policy expressions
• [2009/02/17 10:29] Infinity Linden: granted... it's not the worlds greatest protocol, but it does a very good job of separating policy and mechanism
• [2009/02/17 10:30] Zha Ewry: People will then have the choice of using or not using those mechansisms, but if t hey do want them, they know where they go
• [2009/02/17 10:30] Latha Serevi: Is anybody else but me not interested at all in LLSD and think it's an odd primary focus? :-O
• [2009/02/17 10:30] Hermit Barber: Infinity, your tail is right behind you. If you are quiet and quick you could maybe pounce on it and catch it unawares!
• [2009/02/17 10:31] Rex Cronon: but if u provide for policy expressions that means u don't ignore policy:_
• [2009/02/17 10:31] Infinity Linden: also.. .with respect to security.. if we're talking about host authentication, we simply say... the parties who wish to interoperate will develop a policy for issuing certs and agree to turn cert checking on. this protocol only provides a mechanism for a compliant implementation to use certs for yes/no access decisions
• [2009/02/17 10:31] Morgaine Dinova: We only have a good understanding of how to design a base protocol for SL-type worlds, and what's more, we only know (roughly) how to handle login and part of TP, so we know very little currently. Most importantly, we know nothing about the requirements of non-SL worlds who wish to interop.
• [2009/02/17 10:31] Zha Ewry: Its not a primary focus, other than being low hanging fruit the community would like to look at
• [2009/02/17 10:31] Zha Ewry: And, having the slot for a policy expression, is very different from using it
• [2009/02/17 10:31] Shamir Katsu: @infinity you are making assumptions about implementation
• [2009/02/17 10:32] Zha Ewry: Look at RFC822. Tons of slots which rarely get used, but are in the spec
• [2009/02/17 10:32] Aleric Inglewood: Zha: I hope you don't say that this topic can be delayed (because it cannot); the topic being: Does where and how data is stored and retrieved have a structual effect on the protocol. I hear people say that a protocol can be designed that will work for every possible way one might want to store/retrieve data and for any possible security services that might be needed - but my first hand experience is the opposite: you have to take that into account before the protocol is design or it will turn out to make important changes (ie from central to distributed storage) impossible, or at the very least very non-robust and inefficient in the future.
• [2009/02/17 10:32] Infinity Linden: Morgaine... why don't we let the non-SL worlds speak for themselvs
• [2009/02/17 10:32] Goldie Katsu: yeah certs are nice, but not always the solution, and is only one layer of trust
• [2009/02/17 10:33] Infinity Linden: shudders at the memory of RFC822
• [2009/02/17 10:33] Shamir Katsu: and certs just tell you who you are talking to, not anything about them or what you expect them to do
• [2009/02/17 10:33] Infinity Linden: which is thankfully replaced with RFC1822
• [2009/02/17 10:33] Latha Serevi: Aleric, that's an interesting point. It would be cool to define about four different "use cases" of where the data lives. I still don't quite understand yours.
• [2009/02/17 10:33] Zha Ewry: If you assme the world is largely distributed, then the cases where you centralize subset
• [2009/02/17 10:33] Rex Cronon: i think that alric is right
• [2009/02/17 10:33] Zha Ewry: The reverse is less true
• [2009/02/17 10:34] Zha Ewry: Indeed the UUID is enough assumption in the current Second Life Grid is a strong hint of that
• [2009/02/17 10:34] Infinity Linden: @shamir. yes. i am assuming the implementation implements the defined protocol.
• [2009/02/17 10:34] Infinity Linden: i am assuming that it MAY respond to HTTPS instead of just HTTP
• [2009/02/17 10:34] Morgaine Dinova: Infinity: I certainly hope those other worlds will speak for themselves. But we have to go further than that. Our own goal is to foster broad VW interop, not merely to be forced into it by others under duress. The work has to come from us as well as others.
• [2009/02/17 10:35] Infinity Linden: and that if it responds to HTTPS, the protocol explicitly describes how it may be used to present credentials to the peer
• [2009/02/17 10:35] Infinity Linden: what the peer does with those credentials, i could care less about
• [2009/02/17 10:35] Rex Cronon: https deals with how is data tranported across the web
• [2009/02/17 10:35] Zha Ewry: tosses Infinity's credential into the lake
• [2009/02/17 10:36] Latha Serevi: Question: we all want to see OGP developed. IETF is about highly general plumbing. Are we all going to have to adjust to the fact that the most expansive possible IETF RFC's we would see in the future will only be 1/4 of the OGP story?
• [2009/02/17 10:36] FWord Utorid: meh. protocol schmotocol. this shit should just be in swedish chef UTF. bourgebourgsquare meep 12 bourgebourge
• [2009/02/17 10:36] Shamir Katsu: So we're not talking about something more seperable that allows rules for how you trust another domain to interact with your assets
• [2009/02/17 10:36] Infinity Linden: @Morgaine. no, we dno't hvae to go further than that. it's dangerous to make a protocol where you attempt to define other stakeholder's interests
• [2009/02/17 10:36] Aleric Inglewood: Just thinking about scalabity and the nature of a virtual world, I think we only have to concentrate - but MUST concentrate - on totally DE-centralized storage of data (though that is just my opinion). For example, we are sitting here - we don't care what happens elsewhere - so it makes sense to have all the data related to this place in one spot on the network, and data that is used elsewhere (possibly) on other location in the network.
• [2009/02/17 10:36] Zha Ewry: For an awful lot of what we're likely to end up specifcying, we ought to be able to defer to the http/URI/URL formalisms
• [2009/02/17 10:37] Infinity Linden: @Fword... spoken like the true diplomat you are ;-)
• [2009/02/17 10:37] FWord Utorid: aleric: there should be gravity servers for assets that suck them into the local area
• [2009/02/17 10:37] Infinity Linden: @Shamir... we can, but its not in the base protocol
• [2009/02/17 10:37] Zha Ewry: Interestingly, Aleric I've heard the exact opposite proposed, for equally sound reasons
• [2009/02/17 10:37] Infinity Linden: or
• [2009/02/17 10:37] FWord Utorid: infinity: wtf are you talking about, I am very sensitive and caring
• [2009/02/17 10:37] Latha Serevi: Related to my "IETF is for pluumbing" comment, are we going to have to focus most of our interop (and even standards) efforts in some different forum yet to be found?
• [2009/02/17 10:37] Infinity Linden: the base specification explicitly allows extension
• [2009/02/17 10:37] Zha Ewry: "We sshould permit assets to be cached as widely and as far away from the sims as possible, so we don't end up with hot spots"
• [2009/02/17 10:37] Latha Serevi: I'm also curious whether MPEG-V is part of the mix and how.
• [2009/02/17 10:38] Infinity Linden: @Latha.. there's the AWG, the VWR and MPEG-V
• [2009/02/17 10:38] Zha Ewry: Any MPEG-V implementors are welcome to come and discuss how thier solutions work
• [2009/02/17 10:38] Shamir Katsu: And those extensions are scalable and support different models for interchange between domains?
• [2009/02/17 10:38] Infinity Linden: VWR and MPEG-V are more into "how do we define virtual worlds"
• [2009/02/17 10:39] Infinity Linden: MMOX is.. .given VWs and MMOs as they exist today, what bits are common and can be standardized?
• [2009/02/17 10:39] Morgaine Dinova: Infinity: your own words (and those of Zero) on the existing MMOX documents all point to your desire to make the MMOX protocol generic to VWs, not just to SL and SL clones. I'm taking that as an honest desire, and therefore dismissing your last comment as wishing to consider only your (LL's) interests.
• [2009/02/17 10:39] Infinity Linden: and insert the words "and in the near future" after "as they exist today"
• [2009/02/17 10:39] FWord Utorid: The Court Of Morgaine is now in session
• [2009/02/17 10:39] Morgaine Dinova: purrs
• [2009/02/17 10:40] Zha Ewry: Seriously, I don't see a lot of concrete stuff I can talk about from the other groups
• [2009/02/17 10:40] Morgaine Dinova: Can't have it both ways.
• [2009/02/17 10:40] Latha Serevi: I only heard about MPEG-V yesterday, and I don't know what VWR is. I need help figuring out whwere we should define, say, prims.
• [2009/02/17 10:40] Infinity Linden: why would creating an open protocol that's usable by anyone with input from anyone who wants to show up be a sectional LL interest?
• [2009/02/17 10:40] Aleric Inglewood: Zha: That seems to mean two things: 1) store it away from SIMs (but *centralized*), because 2) the same item(s) might be used in many sims, thus overloading the sim that stores it. In both it is assumed that data is stored centralized, which is what I was fighting. So, what are they proving if they start with the assumption that data will be centralized?
• [2009/02/17 10:40] FWord Utorid: thinks now is the time to make that armpit fart animation and sound effect I'vebeen wanting for so long
• [2009/02/17 10:40] Zha Ewry: No Aleric, you're making a pile of assumptions
• [2009/02/17 10:41] FWord Utorid: aleric: the issue with the decentralized happy is uuid collisions, among other things
• [2009/02/17 10:41] Infinity Linden: hmm... personally.. i don't like to thing about "sims" in the MMOX or OGP context
• [2009/02/17 10:41] Infinity Linden: it's probably better to say "protocol endpoints"
• [2009/02/17 10:41] Zha Ewry: Address things in a way which allow them to be stored, cached and distributed broadly, rather than tied to any single simulation
• [2009/02/17 10:41] FWord Utorid: but really, the problem with the notions of this whole thing are that there's no way to say <prim href="alericassetserver://12/">
• [2009/02/17 10:41] Morgaine Dinova: Aleric: a VW may not even have sims as hosts. It's you who is making assumptions now :-))))
• [2009/02/17 10:41] Infinity Linden: some people will have virtual worlds that are split up into different sims like we do
• [2009/02/17 10:41] Infinity Linden: others will have a different model
• [2009/02/17 10:41] Morgaine Dinova: Indeed Infinity
• [2009/02/17 10:41] Infinity Linden: @Fword... why?
• [2009/02/17 10:42] Zha Ewry: My hope is that the RFCs decribe a set of services which
• [2009/02/17 10:42] Zha Ewry: can be compiosed into a hueg set of depolyment models
• [2009/02/17 10:42] Morgaine Dinova: That's why we'll be talking about comms between endpoints --- and not saying what is at those endpoints.
• [2009/02/17 10:42] Latha Serevi: Aleric, I think you've already won the argument. Assets will NOT be assumed centralized. We just need to figure out how to name them and use them. "just". :-)
• [2009/02/17 10:42] Zha Ewry: And, in general, focus on what needs to bne done to support the endpoint/service not the deployment model
• [2009/02/17 10:42] Morgaine Dinova: Well said Latha.
• [2009/02/17 10:42] FWord Utorid: @Infinity: I like the notion of one contiguous space with little sectors that can be placed on the wasteland. So, there's the world, and then, there's continents. World is all the sphere and the base water and so on, the continents are like image map regions on the globe
• [2009/02/17 10:42] Infinity Linden: @morgaine.. yes.. i make certain assumptions... such as... tcp will be the network protocol... HTTPS will be used... etc.
• [2009/02/17 10:42] Aleric Inglewood: I meantioned sims only because I was reacting to Zha's quote, which refered to sims.
• [2009/02/17 10:43] Infinity Linden: and when i say things like "some other people might not do it the way we do it... i think that's a valid assumption"
• [2009/02/17 10:43] FWord Utorid: @Infinity: alericassetserver://12 is a scheme like html where you can reference items in other grids like a web page references items in other webpages. the lack of that is why this all sucks
• [2009/02/17 10:43] Zha Ewry: +1 FW
• [2009/02/17 10:44] Morgaine Dinova: Latha summed it up fine. I don't think there's disagreement here, only "violent agreement" :-))))
• [2009/02/17 10:44] Zha Ewry: The deep use of UUID is a nightmare, as soon as you habe more than one grid
• [2009/02/17 10:44] Creative Acronym: in truth LL may not wish to do it the way you currently do it forever
• [2009/02/17 10:44] Infinity Linden: FWord... okay... get an RFC for it and we'll promptly create an interoperability profile that says.. "BTW... we ignore alericassetserver: scheme URIs"
• [2009/02/17 10:44] Infinity Linden: but we won't mention we're ignoring it until it's an official URI scheme
• [2009/02/17 10:44] FWord Utorid: @infinity: exactly, which is why the Linden Dinosaur fell before the meteor
• [2009/02/17 10:45] FWord Utorid: and the Mammals walk the planet
• [2009/02/17 10:45] 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 in 15 minutes)
• [2009/02/17 10:45] Infinity Linden: and was replaced by?
• [2009/02/17 10:45] Rex Cronon: using an uuid should be no problem if items are stored i differend databases
• [2009/02/17 10:45] Latha Serevi: I take it that VWR means [1] .... do we know any of those people? Are they in MMOX?
• [2009/02/17 10:45] Zha Ewry: How do you know which database Rex?
• [2009/02/17 10:45] Infinity Linden: @Latha.. yes... we know these people... no they're not MMOX
• [2009/02/17 10:45] Infinity Linden: though
• [2009/02/17 10:45] FWord Utorid: Infinity: it'd be replaced by something dumb like VRML with extensions for physics and anti-suck
• [2009/02/17 10:45] Morgaine Dinova: Creative: indeed, Lindens will be in the same boat as everyone else in MMOX --- the protocols are very unlikely to reflect LL's desired implementation. Interop requires adaptability by everyone concerned.
• [2009/02/17 10:45] Rex Cronon: it has an adress:)
• [2009/02/17 10:45] Infinity Linden: smoe of them are likely to show up at the MMOX meeting and participate
• [2009/02/17 10:45] Zha Ewry: please get me 8736-3789-789fA0789-67FDA now
• [2009/02/17 10:45] Zha Ewry: Oh
• [2009/02/17 10:45] Zha Ewry: Wait..
• [2009/02/17 10:46] Zha Ewry: You have no idea where I stored it
• [2009/02/17 10:46] Infinity Linden: @Morgaine... exactly.. .it's essential that it not ONLY address our sectional interests
• [2009/02/17 10:46] Rex Cronon: tell me where the database resides and i migh get it
• [2009/02/17 10:46] Aleric Inglewood: I think the end result will be several layers of data access: local (fast retrieval, cached), and some way of finding back data from any point (which needs a unique UUID). The only way to get unique UUID's is by generating them by some authority - thus, one layer would have to do with generating new UUIDs (who is responsible for what?), and retrieving data (how to find where data is stored, if not local, once we have a UUID)?
• [2009/02/17 10:46] Infinity Linden: which is why it's being pursued in the IETF
• [2009/02/17 10:46] Zha Ewry: Which is why a naked UUID is useless
• [2009/02/17 10:46] Infinity Linden: but nekked UUIDs are so fetching!
• [2009/02/17 10:46] Shamir Katsu: @zha especially useless without agreements on how Zhaworld will handle objects from Shamirworld
• [2009/02/17 10:46] Zha Ewry: If you're very, very careful, you can probably make good guesses about where a naked UUID should be resolved
• [2009/02/17 10:47] Latha Serevi: I'm still seeking guidance: where shall we define what a standard prim is, when we get around to it? Surely not IETF.
• [2009/02/17 10:47] FWord Utorid: cuddles the sexy UUID, with it's delicate digits
• [2009/02/17 10:47] Zha Ewry: Why not Latha?
• [2009/02/17 10:47] Zha Ewry: Its a low level format that says how to display a bit of formatted data
• [2009/02/17 10:47] Infinity Linden: but yes... we only get away with nekkid UUIDs cause we're a walled garden right now
• [2009/02/17 10:47] Infinity Linden: but
• [2009/02/17 10:47] Creative Acronym: has there been any thought put into a UUID assigning authority?
• [2009/02/17 10:47] Rex Cronon: u could have primary key like: userUUID, itemUUID
• [2009/02/17 10:47] FWord Utorid: anyway, people are going to want the space triangles that they can rez from sweden
• [2009/02/17 10:47] FWord Utorid: creative: ick
• [2009/02/17 10:47] Infinity Linden: persnoally.. i like the idea of using URIs to reference things
• [2009/02/17 10:47] Creative Acronym: don't say ick so fast
• [2009/02/17 10:47] FWord Utorid: uuids should be replaced by uris
• [2009/02/17 10:48] Zha Ewry: Given that UUIDs are inherently supposed to be unique and painfully inscrutable, assigning them is not helpful
• [2009/02/17 10:48] Zha Ewry: URIs is my take too
• [2009/02/17 10:48] Zha Ewry: tho..
• [2009/02/17 10:48] FWord Utorid: creative: why do I want to have to ask for my insanely long number of uniqueness
• [2009/02/17 10:48] Zha Ewry: I saddly observe that URIs are the next great thing and have been for a long time
• [2009/02/17 10:48] Zha Ewry: +! FW
• [2009/02/17 10:48] Infinity Linden: @Zha.. LOL
• [2009/02/17 10:48] Morgaine Dinova: LOL, UUID authority, that's funny. Folks, the whole idea of UUIDs is that they're unique whoever generates them, because of the immensity of the address space. If they're not using, your random number generator is screwed.
• [2009/02/17 10:48] Zha Ewry: thinks about putting +1 FW on a gesture
• [2009/02/17 10:48] FWord Utorid: when proper implementation of the bourgebourge protocol would allow me to have it for free.
• [2009/02/17 10:49] Infinity Linden: well... we had a system like that, it was called Kerberos
• [2009/02/17 10:49] FWord Utorid: wtf zha
• [2009/02/17 10:49] Zha Ewry: makes the sign of the inverted K and holds it high against Kerberos
• [2009/02/17 10:49] Rex Cronon: morgaine, there is always the possiblity of collision:)
• [2009/02/17 10:49] Latha Serevi: Zha - "Prims not in IETF" because (1) IETF is about communication and prims are about VW simulators; (2) IETF track has plenty to work on just authenticating , storing abstract assets, and streaming updates; (3) Standard prims aren't nearly ready.
• [2009/02/17 10:49] Infinity Linden: Fword and Prok are welcome to make a competing protocol developed iwth the ferverent sense of revolutionary zeal
• [2009/02/17 10:49] Infinity Linden: (that was a joke)
• [2009/02/17 10:49] FWord Utorid: infinity: I am already kicking your ass :P
• [2009/02/17 10:50] 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 in 10 minutes)
• [2009/02/17 10:50] Rex Cronon: i doesn't matter how random your random number generator is:)
• [2009/02/17 10:50] Aleric Inglewood: Creative: The problem of a unique ID generating authority is not to generate it (ie: oomx://my_assigned_ID/whatever_is_unique_to_me/), but to find out WHICH authority (my_assigned_ID) is responsible before you generate it :/
• [2009/02/17 10:50] FWord Utorid: zha: how about making a little protocol hack that loads vrml into opensmi?
• [2009/02/17 10:50] FWord Utorid: *sim
• [2009/02/17 10:50] Infinity Linden: or... about prims... we can always define prims later
• [2009/02/17 10:50] FWord Utorid: vrml document to oar converter ftw
• [2009/02/17 10:50] Infinity Linden: but if we don't get the base protocol working, an IETF standard for prims would probably get ignored
• [2009/02/17 10:51] Infinity Linden: and... who knows... maybe we could steal bits of X3D
• [2009/02/17 10:51] Latha Serevi: Assuming that we don't deal with prims at all in IETF, where _else_ shall we do it? I have no clue, does anybody?
• [2009/02/17 10:51] Infinity Linden: X3D, Collada
• [2009/02/17 10:51] Creative Acronym: but think of UUID authority servers like MAC range assignments, you can find the location of an asset's origin and where it is stored by the beginning or ending bytecode
• [2009/02/17 10:51] FWord Utorid: wonders how 3d studio max stores it's prim data and if you could load up a max document and show that shizzle
• [2009/02/17 10:51] Rex Cronon: u mean how ll defines prims?
• [2009/02/17 10:51] Zha Ewry: So.. if you want my full on corporate answer, the obvious thing to do is to get the collada people to do the prim format, and the IETF to ref it.
• [2009/02/17 10:52] Infinity Linden: i'm not terribly happy with the collada prim interaction model (oh heck.. or am i thinking of X3D)
• [2009/02/17 10:52] Zha Ewry: That said? I just want a good way to cannonicalize the prim formats, and allow for easy extension and regsitration. I don't really care who does it
• [2009/02/17 10:52] Morgaine Dinova: Rex: sure, in theory you can get collisions, and also in practice when things are broken. But if you're getting collisions in practice then you'd better fix your code! Of course one does some checks too, but they're not uniqueness checks --- checking UUID uniqueness across interop'ing world is impossible.
• [2009/02/17 10:52] Infinity Linden: makes not to investigate both
• [2009/02/17 10:52] Latha Serevi: Rex, I (hopefully we) are talkign about a standard format for constructive geometry prim-like constructions)
• [2009/02/17 10:52] Infinity Linden: and... also... if we ues HTTP as a transport... we could have MIME types tell us what format the prim data is in
• [2009/02/17 10:52] Latha Serevi: ...and how to limit our expectations of what goes in the IETF track
• [2009/02/17 10:52] Infinity Linden: it's probably not sufficient for true interop
• [2009/02/17 10:53] Zha Ewry: Last I looked, tho, Collada was still a fair way away from a useful framework for constructive geometry and prim sets
• [2009/02/17 10:53] Infinity Linden: you'd still need agreement on which features of each prim format are allowed
• [2009/02/17 10:53] Rex Cronon: there is already a way to store objects, is called an .obj file, and is quite wildly used:)
• [2009/02/17 10:53] Infinity Linden: are you talking about the .obj format that we use in our asset servers?
• [2009/02/17 10:53] Infinity Linden: puhleeeze!
• [2009/02/17 10:54] Latha Serevi: We might need to think about wrapping up our meeting.
• [2009/02/17 10:54] Aleric Inglewood: Rex: I don't consider Microsoft to be "widely" used :p
• [2009/02/17 10:54] Infinity Linden: hmm.. this is weird.. the linden complaining about the linden object format
• [2009/02/17 10:54] FWord Utorid: ... ok, so, we're scrapping the linden mess and making something that would work. so, we'll use the http infrastructure of teh internet
• [2009/02/17 10:54] Zha Ewry: I expect the wavefron OBJ format, Infinity
• [2009/02/17 10:54] Rex Cronon: i am talking about the .obj file used to store 3d objects
• [2009/02/17 10:54] Infinity Linden: @Aleric... i don't think he was talkign about PE or XCOFF
• [2009/02/17 10:54] Aleric Inglewood: Ah - sowwy :p
• [2009/02/17 10:54] Latha Serevi: Anybody have agenda items or "what's next this week" to pass on before we quit for the day?
• [2009/02/17 10:54] FWord Utorid: it'll be in plain text and use the compression happy of triangles needed to shrink the universe down
• [2009/02/17 10:54] Infinity Linden: kk. thx.
• [2009/02/17 10:54] Zha Ewry: its prety much a mesh only format
• [2009/02/17 10:55] 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 in 5 minutes)
• [2009/02/17 10:55] Infinity Linden: well... if you want to make a mesh only virtual world, you should use that format
• [2009/02/17 10:55] Rex Cronon: it also stores other kinds of data
• [2009/02/17 10:55] Zha Ewry: The OBJ file format is a simple data-format that represents 3D geometry alone ‚ namely, the position of each vertex, the texture coordinate associated with a vertex, the normal at each vertex, and the faces that make each polygon.
• [2009/02/17 10:55] Zha Ewry: quotes the quasi reliable wikipedia
• [2009/02/17 10:55] Infinity Linden: mmm... i think we definitely need to add the "model for representation and interaction" document as a BoF or WG work item
• [2009/02/17 10:56] Zha Ewry: But, seriuosly, last I looked, it was a mesh view of geometry
• [2009/02/17 10:56] Morgaine Dinova: Indeed. And we need a tentative list of such topics to be covered by MMOX protocols, at least to prime the pump.
• [2009/02/17 10:56] Infinity Linden: though LL officially says we're not getting NURBS anytime soon. hope springs eternal, and i don't think we should enshrine in standardism any objcet format that cannot repreesnt them
• [2009/02/17 10:56] Rex Cronon: so it becomes a component of the object type u want to make
• [2009/02/17 10:57] Infinity Linden: ahem... we have a list of tentative topics
• [2009/02/17 10:57] Aleric Inglewood: Zha: That would be triangles forming a (possibly closed) 2D surface?
• [2009/02/17 10:57] Zha Ewry: nods
• [2009/02/17 10:57] Latha Serevi: Perhaps MMOX will define how to name, subclass, version, negotiate, and encapsulate the 3-d data formats, but not try to list them or define them?
• [2009/02/17 10:58] Infinity Linden: but i'm beginning to think the likelihood we're in a point to declare consensus around forming a WG after just this upcoming meeting is approaching zero
• [2009/02/17 10:58] Rex Cronon: wg?
• [2009/02/17 10:58] Infinity Linden: @Latha... why would it need to do that?
• [2009/02/17 10:58] Zha Ewry: Working Group
• [2009/02/17 10:58] Morgaine Dinova: We don't need to specify which prims or models or meshes etc etc are usable --- we just need to express an extensible way of allowing people to use what they want.
• [2009/02/17 10:58] Zha Ewry: which is the base unit of work in IETF
• [2009/02/17 10:58] Rex Cronon: ohh
• [2009/02/17 10:58] Infinity Linden: from OGP's perspective, object formats are (or can be) orthogonal
• [2009/02/17 10:59] Infinity Linden: @Morgaine.. right..
• [2009/02/17 10:59] Infinity Linden: but the problem is
• [2009/02/17 10:59] Morgaine Dinova: Latha: indeed!
• [2009/02/17 10:59] Aleric Inglewood: I suppose that "attachment point" will be part of the protocol. Will the protocol allow for attaching to a vertex (or with that precision)? (Asking, because - for example - it's annoying that if you attach a lip piercing to your "chin" and an animation changes the mesh - the piercing just hangs in the air; what is needed here is to tell the client to wish mesh point it is attached).
• [2009/02/17 10:59] Rex Cronon: u can say my obj uses this data format to store the 3dinfo
• [2009/02/17 10:59] FWord Utorid: @infinity: we would have more luck turninginto smurfs with magic powder than we will ever have forming a consensus on how things should work
• [2009/02/17 10:59] Infinity Linden: when you offer eveyone infinite flexibility, you get a world where no one can interoperate cause no one supports exactly the same set of options
• [2009/02/17 10:59] Aleric Inglewood: s/wish/which/
• [2009/02/17 10:59] Infinity Linden: LOL
• [2009/02/17 10:59] Latha Serevi: Infinitiy: surely MMOX can at least help organize the data formats further than "opaque block of data TBD"?
• [2009/02/17 10:59] Zha Ewry: +100 Infinity
• [2009/02/17 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)
• [2009/02/17 11:00] Latha Serevi: For isntance, to the poitn of being able to establish if you and I can speak the same language within an MMOX protocol?
• [2009/02/17 11:00] Infinity Linden: @Latha.. sure... but it's not an issue for a wire protocol
• [2009/02/17 11:00] Zha Ewry: The more optioninoality you have the more it beocmes soup, and you need profiles to determine if you can work together
• [2009/02/17 11:00] FWord Utorid: no longer regards zha's points system as meritorious
• [2009/02/17 11:00] Infinity Linden: it mgiht be useful to consider 802.11
• [2009/02/17 11:00] FWord Utorid: +12000 me
• [2009/02/17 11:00] Dahlia Trimble: would be interested in seeing a formal definition of prim parameters
• [2009/02/17 11:01] Infinity Linden: 802.11 defines a broad range of options and standards
• [2009/02/17 11:01] Infinity Linden: but the "Wi Fi Alliance" created an "interoperability profile" for 802.11 that said... okay... you're gonna use these channels and ignore frequency hopping
• [2009/02/17 11:01] Infinity Linden: (plus a few other options)
• [2009/02/17 11:01] Rex Cronon: dahlia, just look at llgetprimitive parameters:)
• [2009/02/17 11:02] Morgaine Dinova: That;s true of inifinite worlds, because they're sparse with an average density of zero. It's not true of our small possible set of finite worlds though. In our area, we can easily find express generality without finding that worlds can't interop. The set of formats etc is very limited.
• [2009/02/17 11:02] FWord Utorid: yeah, because frying the brains of rats in the lab isn't good for business
• [2009/02/17 11:02] Infinity Linden: i'm thinking that there's no chance we can make a protocol that defines everythign in a concerete manner and is broadly usable
• [2009/02/17 11:02] Dahlia Trimble: Rex, trust me, I've been all through that ;)
• [2009/02/17 11:02] Infinity Linden: so maybe like 802.11 and RTP we could define a framework for a protocol
• [2009/02/17 11:02] Rex Cronon: from that info u can rebuild an objed:)
• [2009/02/17 11:02] Rex Cronon: object*
• [2009/02/17 11:02] FWord Utorid: if you define it in concrete it fails when someone wants jellyfish
• [2009/02/17 11:03] Zha Ewry: I think if we're lucky, we'll end up with a set of services, and formats, and then, some profiles which tell us how to actually get useful grids out of the soup
• [2009/02/17 11:03] Morgaine Dinova: Yep
• [2009/02/17 11:03] FWord Utorid: so you have to ultimately hate someone and not let them have what they want
• [2009/02/17 11:03] Infinity Linden: that says.. this is how worlds communicate the metadata, but interpretation of the data is left as an exercise to teh implementers (and oh yeah.. .here' sthe list of things we know you need to put in an interoperability profile for there to be any hope of peeps interoperating)
• [2009/02/17 11:03] Zha Ewry: Frankly, if we get close to that,w e'll be doing pretty well
• [2009/02/17 11:03] FWord Utorid: I pick gary. We should make one that works the way everyone wants EXCEPT Gary. He doesn't get a say.
• [2009/02/17 11:04] Morgaine Dinova: Also, the "nobody will be able to interop" suggestion fails entirely to note that worlds will pick options to allow themselves to interop.
• [2009/02/17 11:04] Latha Serevi: Are we all cool with continuing this meeting until noon?
• [2009/02/17 11:04] Zha Ewry: I pumkpinize shortly
• [2009/02/17 11:04] Infinity Linden: hmm... do i smell frying rat brains? yum!
• [2009/02/17 11:04] Morgaine Dinova: Or are we going to Opensim?
• [2009/02/17 11:04] FWord Utorid: basically, 'nobody will be able to interop' is the truth unless everyone wants to get along
• [2009/02/17 11:04] Rex Cronon: i have to go to a office hour
• [2009/02/17 11:04] Morgaine Dinova: FWord: andthey do.
• [2009/02/17 11:05] Aleric Inglewood: starts to feel social obligations start nagging.
• [2009/02/17 11:05] 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)
• [2009/02/17 11:05] Zha Ewry: I hope the OpenSim office hour isn't another hour of wimpering about asset stores
• [2009/02/17 11:05] FWord Utorid: so you need a set of circumstances wherein they have to agree in order to survive
• [2009/02/17 11:05] Rex Cronon: bye everybody
• [2009/02/17 11:05] FWord Utorid: Morgaine: the opensim people would probably want to ditch everything in favor of their own thing that is better in their own head
• [2009/02/17 11:05] Morgaine Dinova: Well MMOX is just offering interop --- we can't make the horse drink.
• [2009/02/17 11:05] Zha Ewry: Bye rex
• [2009/02/17 11:05] Dahlia Trimble: has rl... bye all :)
• [2009/02/17 11:05] FWord Utorid: and the LL people would probably want to have something that is better in their own head
• [2009/02/17 11:05] Infinity Linden: if it would help, i'll show up and whimper abotu asset servers that serve 84k cnocurrent users
• [2009/02/17 11:05] Zha Ewry: Heh.
• [2009/02/17 11:05] FWord Utorid: infinity: you can't whimper about problems you cause
• [2009/02/17 11:06] Zha Ewry: That would probably be viewed as bratty Infinity
• [2009/02/17 11:06] Morgaine Dinova: FWord: MMOX will (I believe) integrating in Opensim requirements.
• [2009/02/17 11:06] FWord Utorid: 84k concurrency is because of the happy new client
• [2009/02/17 11:06] FWord Utorid: and the asset servers are your fault too
• [2009/02/17 11:06] Infinity Linden: LOL
• [2009/02/17 11:06] Infinity Linden: it's all our fault... even when it works
• [2009/02/17 11:06] Zha Ewry: And frantic DB spreading
• [2009/02/17 11:06] FWord Utorid: plus cmon, being unhappy about your own success...
• [2009/02/17 11:06] FWord Utorid: and potential job security
• [2009/02/17 11:06] Infinity Linden: hey! don't be dissing my DB schema refactorin!
• [2009/02/17 11:07] FWord Utorid: HOLY CRAP THIS PLACE NEEDS ME
• [2009/02/17 11:07] FWord Utorid: you should be spanked. alot.
• [2009/02/17 11:07] Zha Ewry: OK. On that note, I'm going to go write some code
• [2009/02/17 11:07] Morgaine Dinova: Job security is one things that I doubt anyone needs to worry about in this game :P
• [2009/02/17 11:07] Shamir Katsu: see ya later zha
• [2009/02/17 11:07] Infinity Linden: it was more in the context of ... "don't complain that chain bridge has scaling issues, we all have scaling issues"
• [2009/02/17 11:07] Goldie Katsu: Later Zha
• [2009/02/17 11:07] Morgaine Dinova: Cya Zha --- Opensim now?
• [2009/02/17 11:08] Zha Ewry: I may wander over, but I should actually try to write code
• [2009/02/17 11:08] Infinity Linden: i'm going to go work on docs
• [2009/02/17 11:08] Morgaine Dinova: Or is Opensim at midday?
• [2009/02/17 11:08] Zha Ewry: Feel free to IM me, if they get to something useful
• [2009/02/17 11:08] Zha Ewry: no, its now
• [2009/02/17 11:08] Goldie Katsu: I'm going to go back to my training.
• [2009/02/17 11:08] Infinity Linden: cheers peeps
• [2009/02/17 11:08] FWord Utorid: you should still be spanked.
• [2009/02/17 11:08] Zha Ewry: Laters alls
• [2009/02/17 11:08] Infinity Linden: jedi training?
• [2009/02/17 11:08] FWord Utorid: bye evariwun
• [2009/02/17 11:08] Goldie Katsu: The force is strong in you
• [2009/02/17 11:08] Mirt Tenk: tc
• [2009/02/17 11:08] FWord Utorid: may divorce be with you
• [2009/02/17 11:08] Aleric Inglewood: I think I'm going to leave as well... I think I'm only really interested in UUID generation, and how data is stored and retrieved in a truely scalable way - and we seem to have exhausted the time for that subject anyway today ;)
• [2009/02/17 11:09] Infinity Linden: (spoken on OSGrid): "your god-mode jedi tricks won't work on me!"
• [2009/02/17 11:09] Infinity Linden: Aleric
• [2009/02/17 11:09] Goldie Katsu: These are not the void()'s you're looking for.
• [2009/02/17 11:09] FWord Utorid: sends a SpankAvatarSideways Packet at the sim
• [2009/02/17 11:09] Infinity Linden: aer you interested in that in puerly a theoretical senes
• [2009/02/17 11:09] Infinity Linden: or how it's done in OpenSim
• [2009/02/17 11:09] Morgaine Dinova: Nice dress hermie, but your legs are sticking through ... and that issue may even become relevant in MMOX, depending on our coverage :-)
• [2009/02/17 11:10] Aleric Inglewood: Infinity: Yes, but my motivation is that SL is important to me, I want it to work :).
• [2009/02/17 11:10] Infinity Linden: notes i wear slacks a lot
• [2009/02/17 11:10] FWord Utorid: ... it works now
• [2009/02/17 11:10] Infinity Linden: in RL and SL
• [2009/02/17 11:10] Morgaine Dinova: Hehe Infi :-)
• [2009/02/17 11:10] Infinity Linden: i hate it when my skirts cut through my hips
• [2009/02/17 11:10] Aleric Inglewood: The way it is won't work with 1 million users online.
• [2009/02/17 11:10] Goldie Katsu: And I'm the same way with skirts
• [2009/02/17 11:11] Goldie Katsu: although I can't find as cool flowy outfits as I can here.
• [2009/02/17 11:11] Infinity Linden: @Aleric... yup... which is one of the reasons we're going open with the whole idea
• [2009/02/17 11:11] FWord Utorid: ok. time to go check on my vast financial empire of crime and stuff. remember to take your shirts off
• [2009/02/17 11:11] Goldie Katsu: as many that is. *looks at her coffee cup*
• [2009/02/17 11:12] Infinity Linden: mmm... coffee...
• [2009/02/17 11:12] FWord Utorid: baibai
• [2009/02/17 11:12] Infinity Linden: BAI
• [2009/02/17 11:12] Goldie Katsu: Well next time your in the Denver area we have the best home roasted stuff :)
• [2009/02/17 11:12] Infinity Linden: when's the OSGrid OH?
• [2009/02/17 11:12] Infinity Linden: now or noon?
• [2009/02/17 11:12] Aleric Inglewood: Infinity: The funny things is ... it doesn't have to be truely scalable: one individual has enough for a "life time" at a certain size of a grid. Beyond that we only need a (slow) method for transfering assets to another grid.
• [2009/02/17 11:12] Morgaine Dinova: Think it's noon, lemme check wiki
• [2009/02/17 11:13] Goldie Katsu: Now on my gcal
• [2009/02/17 11:13] Goldie Katsu: ok that training thing.
• [2009/02/17 11:13] Aleric Inglewood: Ie, 100 grids of 100,000 online people each will do fine - as long as you can TP between a few, and copy your assets for the most parts.
• [2009/02/17 11:14] Aleric Inglewood: So, it doesn't have to be mathematically scalable.... We can solve it with just a few layers: fast, intermediate, and one final "seldom accessed" central authority.
• [2009/02/17 11:15] Morgaine Dinova: Aleric, that statically tiled resource approach is inherently non-scalable, since your resources don't follow where your load is. So we're certainly not hardwiring that architecture in.
• [2009/02/17 11:16] Morgaine Dinova: AW Groupies page says it's midday in Wright Plaza
• [2009/02/17 11:17] Infinity Linden: but... tiled resource has one thing going for it... it's easy to bill. or it's a lot easier to communicate what you get when you pay your money
• [2009/02/17 11:17] Aleric Inglewood: Sorry, I can't follow - perhaps there is some misunderstanding :/. But well... as Zha stated, we're not here to solve the details anyway...
• [2009/02/17 11:18] Infinity Linden: personally.. speaking not as a linden here... i see a future of what bruce sterling calls "suizery" or "feudalism" where some core ICANN like body divvies out control to grant titles to a number of registrars
• [2009/02/17 11:18] Morgaine Dinova: Infinity: yeah, but MMOX isn't being driven by ease to bill :P. A provider can certainly use static resource tiling, but we just don't hardwire it in to the mechanism, as it's an expression of policy.
• [2009/02/17 11:18] Aleric Inglewood: What i have in mind is very abstract and mathematical... Normally people never understand me :/ ... Perhaps I should not get involved at all, as it would very likely more time than I want to make free for this to get any ideas I might have accross
• [2009/02/17 11:18] Infinity Linden: and what you wind up paying for is the maintenance of the network directory taht says your servers are authoratative for a portion of the virtual land
• [2009/02/17 11:18] Morgaine Dinova: Aleric: I'm interested :-)
• [2009/02/17 11:19] Aleric Inglewood: phone..
• [2009/02/17 11:19] Infinity Linden: me too
• [2009/02/17 11:19] Infinity Linden: but keep in mind... the MMOX / OGP effort is one to define a protocol, not to investigate the problem domain of virtual worlds
• [2009/02/17 11:19] Aleric Inglewood: back
• [2009/02/17 11:19] Aleric Inglewood: Ok... the meeting is over?
• [2009/02/17 11:19] Infinity Linden: yup
• [2009/02/17 11:20] Infinity Linden: i'm gonna go type 'til twelve