User:Zero Linden/Office Hours/2009 apr 07

[2009/04/07 12:58] Gypsy Paz: hey Zero
[2009/04/07 12:58] Zero Linden: hello
[2009/04/07 12:58] Wyn Galbraith: Afternoon Zero :)
[2009/04/07 12:58] Gypsy Paz: you lost some weight zero ;)
[2009/04/07 12:58] Moon Metty: hi Zero :)
[2009/04/07 12:58] Morgaine Dinova: Hiya Zero :-)
[2009/04/07 12:58] Zero Linden: The IETF will do that to ya...
[2009/04/07 12:58] Zha Ewry: Afternoon all
[2009/04/07 12:58] Gypsy Paz: ;)
[2009/04/07 12:59] Morgaine Dinova: Hahahaha
[2009/04/07 12:59] Zha Ewry: Now, now, Zero, it's all optional weight
[2009/04/07 12:59] Wyn Galbraith: Optional weight?
[2009/04/07 12:59] Zha Ewry: Sure
[2009/04/07 12:59] Wyn Galbraith: hrms and falls silent :D
[2009/04/07 12:59] Zha Ewry: By specing only what you cana gree on, an making everything else optional
[2009/04/07 12:59] Zha Ewry: The weight gets very small
http://slurl.com/secondlife/Grasmere/163/111/27 (Starts now)
[2009/04/07 13:00] Zha Ewry: gestures one more time to Morgaine and the beanbag parked next to her
[2009/04/07 13:00] Aimee Trescothick: the more you tell a cat to do something ...
[2009/04/07 13:00] Zha Ewry: Yeah, really
[2009/04/07 13:00] Aimee Trescothick: :D
[2009/04/07 13:00] Wyn Galbraith: Correct
[2009/04/07 13:01] Zero Linden: We'll wait a few more minutes here...
[2009/04/07 13:01] JayR Cela: ooo ooo city liter needs changing / :_)
[2009/04/07 13:01] Zha Ewry: The new kitty totally believes in one way delegation. "Hey, Human, more food, please"
[2009/04/07 13:01] Annie Obscure: city litter in my city sure needs it
[2009/04/07 13:01] Geo Meek: .
[2009/04/07 13:01] Zha Ewry: "What, what? You want *me* to do something?"
[2009/04/07 13:01] Geo Meek: that too
[2009/04/07 13:01] Morgaine Dinova: Zero --- agenda items :-) [1]
[2009/04/07 13:01] Zha Ewry: "I think you misunderstand this relationship"
[2009/04/07 13:01] Zero Linden: Yes - I've seen the list
[2009/04/07 13:02] Zha Ewry: Oooh. The beanbag chair looks good on FIC
[2009/04/07 13:02] Zero Linden: I'm going to put one before those: IETF/MMOX report/debrief/counsiling/etc....
[2009/04/07 13:02] Zha Ewry: doesn't have enough mental health coverage to pay for IETF counsuling
[2009/04/07 13:02] FIC Messenger: :D
[2009/04/07 13:02] Morgaine Dinova: The BoF was traumatic? ;-)
[2009/04/07 13:02] Morgaine Dinova: I thought it went very well, hehe
[2009/04/07 13:02] JayR Cela: I would just like to be able to take my own stuff onto the open simm grid
[2009/04/07 13:03] JayR Cela: thats all i want / no fuss / no muss :_)
[2009/04/07 13:03] Geo Meek: all things here belong to LL
[2009/04/07 13:03] Zero Linden: well .. you weren't in the room - I was in the direct line of fire - shrapnel everywhere
[2009/04/07 13:03] Zha Ewry: Oh, it wasn't bad
[2009/04/07 13:03] Morgaine Dinova: Geo: No they don't. They very expressly belong to their creators. It says so in black and white.
[2009/04/07 13:04] JayR Cela: ok / look at it this way....
[2009/04/07 13:04] JayR Cela: LL rents us land .......
[2009/04/07 13:04] Geo Meek: so you say
[2009/04/07 13:04] Zero Linden: "All your prims are belong to LL"?
[2009/04/07 13:04] Wyn Galbraith: server space
[2009/04/07 13:04] JayR Cela: the stuff i create is mine
[2009/04/07 13:04] JayR Cela: i want to be able to take it with me
[2009/04/07 13:04] Zha Ewry: If you create it, you have co-copyright. If you include anyone elses content, you have to talk to them
http://slurl.com/secondlife/Grasmere/163/111/27 (Started 5 minutes ago)
[2009/04/07 13:05] Wyn Galbraith: the content is ours the server space (land) is LLs
[2009/04/07 13:05] Zero Linden: yes - note: My quote was a joke
[2009/04/07 13:05] Zero Linden: okay
[2009/04/07 13:05] Zero Linden: 5 min after
[2009/04/07 13:05] Zero Linden: let's begin
[2009/04/07 13:05] JayR Cela: ok
[2009/04/07 13:05] Wyn Galbraith: LOL Zero ;)
[2009/04/07 13:05] Geo Meek: please
[2009/04/07 13:05] Zero Linden: The agenda so far is
[2009/04/07 13:05] Zero Linden: 1) IETF/MMOX debrief
[2009/04/07 13:05] Zha Ewry: really needs a Rhino Ave.. After the BOF, I clearly have Rhino skin
[2009/04/07 13:06] Zero Linden: 2) AWG communications
[2009/04/07 13:06] Zero Linden: 2) MMOX/OGP design discussions
[2009/04/07 13:06] Zero Linden: 3) Next for OGP
[2009/04/07 13:06] Zero Linden: (sorry about the mis numbering)
[2009/04/07 13:06] Zero Linden: anything else?
[2009/04/07 13:06] Zero Linden: going
[2009/04/07 13:07] Zero Linden: oing
[2009/04/07 13:07] Zha Ewry: Does 2) include Morgaine's item?
[2009/04/07 13:07] Zha Ewry: (s)
[2009/04/07 13:07] JayR Cela: inter-op
[2009/04/07 13:07] Morgaine Dinova: 4) is wrong. 4) is Decoupling asset storage.
[2009/04/07 13:07] Zero Linden: the last three are some kind of soooper short summary of morgains
[2009/04/07 13:07] JayR Cela: TP to differnt OS grids
[2009/04/07 13:07] Zha Ewry: k
[2009/04/07 13:07] Zero Linden: well - I have to say, I didn't get that from the topic
[2009/04/07 13:07] Zero Linden: but - okay
[2009/04/07 13:07] Zero Linden: BUT -
[2009/04/07 13:07] JayR Cela: with same skin and clothing
[2009/04/07 13:08] Zero Linden: that is a rather technical topic to discuss w/o any written proposal....
[2009/04/07 13:08] Zero Linden: but -when we get there....
[2009/04/07 13:08] Zero Linden: SO
[2009/04/07 13:08] Zero Linden: 1) IETF/MMOX
[2009/04/07 13:08] Zero Linden: I wish Infinity were here to summarize
[2009/04/07 13:08] Zero Linden: were most people there/on-Jabber/read-the-transcript?
[2009/04/07 13:08] Zero Linden: or do we need a re-cap?
[2009/04/07 13:09] Zero Linden: short re-cap
[2009/04/07 13:09] Zha Ewry: "It was loud, it was full of people who didn't do thier homework"
[2009/04/07 13:09] JayR Cela: as it is right now / totally worthless to TP or enter an OS grid / because i look like a noob
[2009/04/07 13:09] Morgaine Dinova: No time for recaps in these short meetings. Being up to date should be assumed.
[2009/04/07 13:09] JayR Cela: so why bother
[2009/04/07 13:10] Latif Khalifa: i was listening on audio link, it was a very confusing meeting without much in a way of agenda imho :)
[2009/04/07 13:10] Morgaine Dinova: JayR: your interest will be in agenda item 4) then :-)
[2009/04/07 13:10] Zero Linden: well -
[2009/04/07 13:11] Zero Linden: there was (and is still) a published agenda, that followed pretty much to the minute
[2009/04/07 13:11] JayR Cela: Morgaine / OK / kool / I get with you later
[2009/04/07 13:11] Zero Linden: basically here's the deal
[2009/04/07 13:11] Latif Khalifa: so Zero if you could give us your impressions of it in a couple of lines it would be much appreciated
[2009/04/07 13:11] Zero Linden: 85 people showed up
[2009/04/07 13:11] Zero Linden: that is pretty big, and I lost the bet that we wouldn't draw more than 50
[2009/04/07 13:11] Zero Linden: (mind you OAuth's BoF drew well over 100, perhaps 200!)
[2009/04/07 13:12] Zero Linden: (on the other hand HTTPbis drew like 25 people tops)
[2009/04/07 13:12] Zero Linden: So - there is interest
[2009/04/07 13:12] Zero Linden: we had three sets of presentations - those with submitted drafts
[2009/04/07 13:12] Zero Linden: spoke about their drafts - though mostly about the motivations and intentions
[2009/04/07 13:12] Zero Linden: rather than the technical details
[2009/04/07 13:13] Zero Linden: some IETF folks (no MMOX or VW folks) expressed furstration with one or more of:
[2009/04/07 13:13] Zero Linden: a) lack of clear group concensus within MMOX
[2009/04/07 13:13] Zero Linden: b) lack of very concrete objectives
[2009/04/07 13:13] Zero Linden: c) lack of specific protocol to be developed
[2009/04/07 13:13] Morgaine Dinova: JayR, those two links in "Topic 3" in the agenda give you the background, discussed extensively on MMOX list.
[2009/04/07 13:14] Zero Linden: d) totally new language that none of them understood
[2009/04/07 13:14] Dahlia Trimble: protocol to be developed?
[2009/04/07 13:14] Zero Linden: correct: "What protocol are you developing?"
[2009/04/07 13:14] Zero Linden: as in "what specific sets of bytes are you planning to define for transfer over the wires ?"
[2009/04/07 13:15] Larry Massey: the ietf doesn't like to specify user agent behavior
[2009/04/07 13:15] Latif Khalifa: and if its client - server protocol or world - world protocol
[2009/04/07 13:15] Dahlia Trimble: I would have thought they would want a protocol that had been developed and implemented
[2009/04/07 13:15] Morgaine Dinova: Gotta remember it was just a BoF. By the IETF's own rules, BoFs are meant for examining the problem space, and possibly charter, Any solutions would be merely examples, at the BoF stage.
[2009/04/07 13:15] Zero Linden: well - let's be clear, they LOVE to specify the behavior of how the client should send/receive data
[2009/04/07 13:16] Zha Ewry: They want very concret notions of exactly what you will be specifying
[2009/04/07 13:16] Zero Linden: Correct, Morgaine
[2009/04/07 13:16] JayR Cela: OK / every one should read this [....http://www.exitreality.com/]
[2009/04/07 13:16] Zha Ewry: Ideally at a level where they can whine about the order of your commas
[2009/04/07 13:16] Larry Massey: there is no "they" there
[2009/04/07 13:16] Morgaine Dinova: Zha: much later, not at the BoF
[2009/04/07 13:16] Gypsy Paz: exitreality uses very old VRML technology
[2009/04/07 13:16] Zero Linden: I actually went to a session where commas in RFCs was discussed
[2009/04/07 13:16] Zero Linden: !
[2009/04/07 13:16] JayR Cela: 3d worlds in a browser
[2009/04/07 13:17] Zero Linden: SO
[2009/04/07 13:17] Zha Ewry: /Not so much later, Morgaine. They were pretty much wanting a list of specs
[2009/04/07 13:17] Zero Linden: after the exasperation was done with at the BoF
[2009/04/07 13:17] JayR Cela: read the web page
[2009/04/07 13:17] Zero Linden: the big thing came when the Area Director asked for a show of hands if there was interest in persuing a solution in the OGP space
[2009/04/07 13:18] Zero Linden: dozen hands
[2009/04/07 13:18] Zero Linden: (that's good)
[2009/04/07 13:18] JayR Cela: allready Quake 3 is available for download as a plugin for firefox
[2009/04/07 13:18] Zero Linden: So - the big take away I think is
[2009/04/07 13:18] JayR Cela: so how abot SL going the same route?
[2009/04/07 13:19] Larry Massey: is Lisa here? :)
[2009/04/07 13:19] Zero Linden: Come back with a core committed to a particular architecture (OGP), define your needs and objectives and protocols to implement clearly, and start small
[2009/04/07 13:19] JayR Cela: at least look into the possibilities
[2009/04/07 13:20] Latif Khalifa: Zero, the outcome sounds ok then
[2009/04/07 13:20] Zero Linden: so - next IETF is in July, in Sweden
[2009/04/07 13:20] Zero Linden: Latif - yes, I think the outcome was fine -
[2009/04/07 13:20] Larry Massey: you don't need to wait for the meeting
[2009/04/07 13:20] Zero Linden: at least we got asked back!
[2009/04/07 13:20] Zero Linden: :-)
[2009/04/07 13:20] Latif Khalifa: hehe
[2009/04/07 13:20] Zero Linden: No Larry, we won't be
[2009/04/07 13:21] Zha Ewry: Hardly, in fact, getting stuff done, prior to the next meeting is clearly vital
[2009/04/07 13:21] Zero Linden: Infinity will be working on some new drafts to help this
[2009/04/07 13:21] Zero Linden: along
[2009/04/07 13:21] Zero Linden: I'm hope some of you will be too
[2009/04/07 13:21] Morgaine Dinova: Zero: I think the reason why you got only a small fraction interested in the OGP direction is because it hasn't been presented as an open framework which could be used to embrace a huge diversity of worlds. People have been very worried that OGP is nothing more than LL wanting SL everywhere, otherwise you'd have got more pro votes.
[2009/04/07 13:21] Zha Ewry: Now that they've unclogged the mailing list, I'm goign to take a second pass at charter bashing, with a nicely narrow focus
[2009/04/07 13:21] Zero Linden: Well, I suppose I see that from the totally other side of the mirror
[2009/04/07 13:21] Zero Linden: I think 12 votes was huge (!)
[2009/04/07 13:22] Zero Linden: given the set up
[2009/04/07 13:22] Larry Massey: I thought the problem with OGP was that it wasn't clear what problem it was trying to solve -- if you didn't already know before you sat down to look at it
[2009/04/07 13:22] Zha Ewry: Right, anyone who didin't at least make the Area meeting, was lost
[2009/04/07 13:22] Zero Linden: and I think the feedback and advice at the BoF was DONT make this an all emcompassing framework
[2009/04/07 13:22] Zha Ewry: And. a lot of people walked in cold
[2009/04/07 13:22] Zha Ewry: nods
[2009/04/07 13:23] Zha Ewry: The "You keep saying you don't mean to boil the ocean, prove it to us" chant was loud
[2009/04/07 13:23] Zero Linden: The concern that this was just LL pushing SL came from very very very few people, and none of the IETF folks suggested so
[2009/04/07 13:23] Larry Massey: people really want to solve the really hard problem of inter-world interoperability, but that's so hard, something smaller could actually be accomplished
[2009/04/07 13:24] Morgaine Dinova: Zero: but if you don't make it a broad framework, then you can't call the protocol "MMOX", because breadth is implied in the name. :-) You made a rod for your own back ^_^
[2009/04/07 13:24] Larry Massey: don't get stuck on the name
[2009/04/07 13:24] Zero Linden: I'm not bound by acronyms
[2009/04/07 13:24] Zha Ewry: is perfectly happy to call it swimming Pickles
[2009/04/07 13:24] JayR Cela: has to be complete open source
[2009/04/07 13:24] Larry Massey: well, there might be more than one "it"
[2009/04/07 13:24] Zero Linden: or at least, I try not to be!
[2009/04/07 13:24] Zha Ewry: Names, really, are the least of the issues
[2009/04/07 13:24] Dahlia Trimble: trademarks swimming pickles...
[2009/04/07 13:24] Larry Massey: there's a big difference between open source and open standards
[2009/04/07 13:25] Morgaine Dinova: If you're made an OGP workgroup, we'd have some months of solid work behind us by now :-)
[2009/04/07 13:25] Morgaine Dinova: you'd*
[2009/04/07 13:25] Zha Ewry: Umm. The specs submitted were, 90% OGP
[2009/04/07 13:25] Larry Massey: the IETF might not want to provide meeting space if the group has too narrow a focus
[2009/04/07 13:25] Morgaine Dinova: The name has made all the difference.
[2009/04/07 13:25] Zha Ewry: But.. the call was broad enough to make sure other people got involved
[2009/04/07 13:25] Zero Linden: which does represent solid work
[2009/04/07 13:26] Zero Linden: Larry - I don't see that as a risk... have you looked at the charter of some working groups?
[2009/04/07 13:26] Zero Linden: :-)
[2009/04/07 13:26] Zha Ewry: refuses to comment on the charters which inolve adding 3 commas and two clairifyin notes to a single spec
[2009/04/07 13:26] Zha Ewry: (ok, they aren't that bad, but we're fine)
[2009/04/07 13:27] Morgaine Dinova: It's perfectly fine to have a narrow workgroup. But you don't work on one very narrow and specific solution, while portraying it as a very broad and all-encompassing one. :-)
[2009/04/07 13:27] Zero Linden: The key will be - what is the protocol, and what are the group of interested people, that being in the IETF will help
[2009/04/07 13:28] Larry Massey: sorry, i meant narrow community of interest. Narrow focus is good
[2009/04/07 13:28] Zero Linden: if you can define the protocol, and define who's involved, and show where the standardization helps --- IETF will be fine with you
[2009/04/07 13:28] Larry Massey: well, it seemed to me that there were several interoperability issues that could be decoupled, and picking a few to start with and nailing them would be really helpful
[2009/04/07 13:28] Zero Linden: agreed
[2009/04/07 13:29] JayR Cela: byeee every one RL calling
[2009/04/07 13:29] Morgaine Dinova: Zha introduced a nice term today at Groupies, "slots", to make us think more concretely about the protocol being extensible. That kind of approach will be useful I think.
[2009/04/07 13:30] Zha Ewry: (Well, people seem to be having a wicked hard time thinking about how extensability/optionality works in things like LLSD marked up protocols)
[2009/04/07 13:30] Larry Massey: and some things might be data formats that aren't exactly 'protocols'
[2009/04/07 13:30] Morgaine Dinova: 30 mins gone, we're still on #1
[2009/04/07 13:31] Zero Linden: #1 was important
[2009/04/07 13:31] Zero Linden: but
[2009/04/07 13:31] Zero Linden: does look like we are coming to a lull here
[2009/04/07 13:31] Zero Linden: so - any last questions comments about the IETF/MMOX experience?
[2009/04/07 13:31] Morgaine Dinova: The result of the BoF was "Go back to mailing list", as expected, so that's the conclusion.
[2009/04/07 13:32] Techwolf Lupindo: My observation of protucules is that ones that are exteniable tend to last the longest.
[2009/04/07 13:32] Larry Massey: i think the challenge is to come to a consensus about a charter
[2009/04/07 13:32] Larry Massey: the next BoF if there is one should be to nail down the charter
[2009/04/07 13:32] Morgaine Dinova: Did the ADs actually say anything about the prospects of getting a WG approved?
[2009/04/07 13:32] Zha Ewry: Overall, I'd argue it was favoribale, if we get more focus
[2009/04/07 13:33] Zha Ewry: With the caveat, that the number of people who said "That went really well" after the BOF scares me.
[2009/04/07 13:33] Larry Massey: you get a WG if you can get rough consensus on a charter, and people willing to work on it
[2009/04/07 13:33] Zha Ewry: Because, that's pretty scary
[2009/04/07 13:33] Larry Massey: and credible goals
[2009/04/07 13:33] Zero Linden: it was clear that there was no charter for a WG at this time
[2009/04/07 13:33] Zha Ewry: Right, but also clear that there was suport for a second pass at one, with more focus
[2009/04/07 13:34] Zero Linden: true - at least half a dozen IETF folks told me that went well!
[2009/04/07 13:34] Morgaine Dinova: Well if you take the *expectations* for the BoF to be those given (very nicely) by Infinity, then I think it's clear it was a major success.
[2009/04/07 13:34] Zha Ewry: chuckles
[2009/04/07 13:34] Morgaine Dinova: Since it fulfilled its expectations
[2009/04/07 13:34] Zero Linden: okayi
[2009/04/07 13:34] Zha Ewry: As in "Not actually killed, and cut into chunks on th floor of the conference room"
[2009/04/07 13:34] Larry Massey: scope, deliverables, timeline, all the normal stuff in an IETF working group charter
[2009/04/07 13:35] Zero Linden: morgaine placed three topics on the agenda
[2009/04/07 13:35] Zha Ewry: All of those were in the first draftc harter
[2009/04/07 13:35] Morgaine Dinova: How much time do we have today Zoer?
[2009/04/07 13:35] Morgaine Dinova: Zero*
[2009/04/07 13:36] Zero Linden: #1 looks like a statement: "With reference to: LL+AWG/OGP materials / OGP Teleport-5 / 3-world OGP interop scenario
Discussion re non-local teleport and TP scalability."
[2009/04/07 13:36] Zero Linden: we have as much time as we need
[2009/04/07 13:36] Zero Linden: I can go three hours if need by
[2009/04/07 13:36] Zero Linden: but I doubt we'll need that
[2009/04/07 13:36] Zero Linden: well - sorry, the P.S. in #1 is a statement
[2009/04/07 13:36] Larry Massey: is there an updated charter that takes into account the discussion so far?
[2009/04/07 13:36] Larry Massey: charter proposal
[2009/04/07 13:36] Morgaine Dinova: Coolness. You may have noticed that I made a remark there about this 1 meeting per month being totally inadequate for addressing AWG/OGP technical issues.
[2009/04/07 13:37] Zero Linden: Larry - I belive Infinity is going to write a new one in a month or so
[2009/04/07 13:37] Zero Linden: I'm sure others will write some
[2009/04/07 13:37] Zero Linden: yes, - I saw the remark - I disagree with it
[2009/04/07 13:37] Larry Massey: well, focus on the document. it's easy to get sandbagged by theoretical discussions about hypothetical charters
[2009/04/07 13:38] Zero Linden: Exactly!
[2009/04/07 13:38] Zero Linden: in that vein
[2009/04/07 13:38] Larry Massey: and the open issues around the charter -- what are he questions that need to be resolved? What things do everyone agree on
[2009/04/07 13:38] Larry Massey: send out a new draft charter every 2 weeks
[2009/04/07 13:38] Zero Linden: I see the teleport document references, and the 3-way scenario e-mail
[2009/04/07 13:38] Morgaine Dinova: This is our (simplified) interop scenario: [2] The tech issue I'd like to address is how we're going to do the teleporting in that scenario, given that worlds A and C don't know about each other.
[2009/04/07 13:39] Zero Linden: "Discussion re non-local teleport and TP scalability."
[2009/04/07 13:39] Morgaine Dinova: Yep
[2009/04/07 13:39] Larry Massey: well, it seems like it needs an active proxy to map between world models
[2009/04/07 13:39] Zha Ewry: notes "our" is a ill defined term
[2009/04/07 13:39] Larry Massey: and active proxies don't scale. am i misunderstanding this?
[2009/04/07 13:39] Larry Massey: it's like web transcoding proxies: one world needs their metrics, models, physics mapped to the vocabulary of the others
[2009/04/07 13:40] Zero Linden: Uhm -
[2009/04/07 13:40] Zero Linden: actually - I dont' even understand, in the context of OGP, what it means to say "A and C dont' know about eachother"
[2009/04/07 13:40] Zero Linden: in the context of OGP - the formulation is unclear
[2009/04/07 13:40] Zero Linden: A isn't a "virtual world" it is both an agent domain and a region domain
[2009/04/07 13:41] Larry Massey: i don't understand that, can you expand?
[2009/04/07 13:41] Zero Linden: as the user teleports to B and C, it is A's Agent Domain that is interacting with B's and C's region domains
[2009/04/07 13:41] Morgaine Dinova: Zero: numerous posts in the list have discussed TP, and we use it rather loosely, although I tightened it up for Jon Watte. However, we've never dealt with the issue I mentioned above re TP, and there's the scalability issue related to not being able to have trust agreements with millions of worlds too, to which Infinity agreed.
[2009/04/07 13:41] Zero Linden: I don't see how "A and C don't know about each other"
[2009/04/07 13:41] Morgaine Dinova: A has no knowledge of C, at least initially.
[2009/04/07 13:41] Zero Linden: and no, there is no active proxy involved
[2009/04/07 13:42] Larry Massey: i think you're better off using some other word than 'know'
[2009/04/07 13:42] Zero Linden: except that the user followed a link toa region in C?
[2009/04/07 13:42] Morgaine Dinova: Remember that there are a million C's.
[2009/04/07 13:42] Zero Linden: but then A has had to open a connection to C, and -- well , introduce each other, and decided what to exchange, yes?
[2009/04/07 13:42] Zha Ewry: Well, the web definitino of C" is URL/URI, know?
[2009/04/07 13:42] Larry Massey: what do you mean by 'agent domain'?
[2009/04/07 13:43] Morgaine Dinova: No, the user TP'd to B, with which you have a trust agreement, then walked over to C, because B and C know about each other. A doesn't know about C, at least initially
[2009/04/07 13:43] Harleen Gretzky: A would never be able to get to C if his AD does not know about C, regardless of B/C;s relationship
[2009/04/07 13:43] Zero Linden: you can't walk over to C
[2009/04/07 13:43] Mystical Demina: i think the trust model could be the server and the client, and it is the responsiblitiy of the client to provide the adress and any tokens that enable security, but there doesn't be to be agreed standard from the server to retreive any items for the endpointsthe client provide
[2009/04/07 13:43] Zero Linden: regions that belong to separate region domains can't be (virtually) physically adjacent
[2009/04/07 13:44] Zero Linden: though, even if they were --- this form of region transfer is nothing less than a teleport w/o the visual effect
[2009/04/07 13:44] Morgaine Dinova: Zero: that's the problem then. There is an agreement for interop between worlds B and C. World A has no right to deny that interop.
[2009/04/07 13:44] Larry Massey: I'm still trying to understand whether you're talking about the "same model" or "different model" case
[2009/04/07 13:44] Morgaine Dinova: World A is merely providing identity.
[2009/04/07 13:44] Zero Linden: not sure what "model" refers to --- but I think we are discussing OGP teleport -
Preview Grid (Aditi) (Starts in 15 minutes)
[2009/04/07 13:45] Zero Linden: so the worlds have a basic common architecture: Agent Domains, Region Domains, Agents/Avatars as user instantiation
[2009/04/07 13:45] Larry Massey: you have two worlds, one is 3D and the other is 2D
[2009/04/07 13:45] Morgaine Dinova: Larry: this is for OGP in MMOX interop, it's not SL-only OGP.
[2009/04/07 13:45] Zero Linden: makes no differene, Larry
[2009/04/07 13:45] Larry Massey: so if I have a 2D avatar and I go into a 3D world, what do I look like?
[2009/04/07 13:45] Zero Linden: so long as the 2D world is willing to accept an OGP teleport protocol from the agent domian in the 3D world
[2009/04/07 13:45] Larry Massey: and who does the transcoding?
[2009/04/07 13:46] Wyn Galbraith: a well rounded 2D
[2009/04/07 13:46] Zero Linden: there is no transcoding
[2009/04/07 13:46] Zero Linden: we haven't even gotten to the point at
[2009/04/07 13:46] Zero Linden: which the viewer's do
[2009/04/07 13:46] Morgaine Dinova: Larry: that's a mapping problem, lots of methods. It's secondary here, not a showstopper.
[2009/04/07 13:46] Zero Linden: content negotiation for what formats they understnad
[2009/04/07 13:46] Zero Linden: BUT
[2009/04/07 13:46] Mystical Demina: seems if you can't provide a 2D appearance you iwill get one assiend or be given a change to selelct one
[2009/04/07 13:46] Zero Linden: bascially, 2D = web, 3D will equal some minimal set of 3D, and optional extras
[2009/04/07 13:47] Wyn Galbraith: Universal Ruth
[2009/04/07 13:47] Saijanai Kuhnri: Ashton: friendly greetings!
[2009/04/07 13:47] Larry Massey: well, two different 3D models, with different skeltons, articulation points, layering, etc.
[2009/04/07 13:47] Zero Linden: frankly, I don't think the "model" issue , atht elevel of "my graphics are nurbs" vs. "my graphics are CG prims" is really an issue
[2009/04/07 13:47] Larry Massey: 2D and 3D was just an example
[2009/04/07 13:47] Zero Linden: right -
[2009/04/07 13:47] Larry Massey: you say it's not an issue, but it seems critical path to me
[2009/04/07 13:47] Morgaine Dinova: Zero: not quite. VW's are "Any-D" ... it's an implementation detail, the "D'ness". For example, a chat client could enter a world as zero-D.
[2009/04/07 13:47] Larry Massey: for getting 'teleport'
[2009/04/07 13:47] Zero Linden: what the base of OGP is trying to do is
[2009/04/07 13:47] Zha Ewry: The region, and the client, are giong to end up with a content negotiation level
[2009/04/07 13:48] Larry Massey: well, before you can do 'content negotiation' there acutally has to be content
[2009/04/07 13:48] Zero Linden: manage the appearance of the user's agent within the location of the other region
[2009/04/07 13:48] Zha Ewry: which is going to fail, horribly, if the client has no way to handle the content of the new region
[2009/04/07 13:48] Larry Massey: so either the content is there, or else someone is doing the translation
[2009/04/07 13:48] Zero Linden: Larry - think of it this way
[2009/04/07 13:48] Zha Ewry: (tho, is can ahnd off to one which does)
[2009/04/07 13:48] Larry Massey: and if you're doing the translation on the fly, that's an active proxy
[2009/04/07 13:48] Wyn Galbraith: It seems to me that first we have to get there before we can even think about how we look when we get there.
[2009/04/07 13:48] Zero Linden: OAuth mediates access tokens from one web site to another
[2009/04/07 13:48] Zero Linden: but says NOTHING about what those tokens are for, or how they are encoded, or how they are used
[2009/04/07 13:48] Morgaine Dinova: waits patiently for the discussion to return to TP and 3 worlds :-)
[2009/04/07 13:48] Larry Massey: well, if you 'get there' but you can't talk and don't have an appearance, then you're not there
[2009/04/07 13:48] Zero Linden: OGP mediates the appearance of an agent in a region
[2009/04/07 13:49] Zero Linden: at the lowest level
[2009/04/07 13:49] Larry Massey: I'm talking about what's critical path for TP and 3 worlds
[2009/04/07 13:49] Zha Ewry: Right, Larry, but. webishly, that's about the same as getting a link to a content type you can't render on your browser
[2009/04/07 13:49] Larry Massey: i just am stuck on an assumption you're making that i don't understand
[2009/04/07 13:49] Zha Ewry: You "get there"
[2009/04/07 13:49] Harleen Gretzky: If A's AD has no relationship witch C's RD how can A get to C in the first place?
[2009/04/07 13:49] Zero Linden: well - with OAuth - if you "get the token" but don't know what to do with it-- same issue, no?
[2009/04/07 13:49] Zero Linden: BUT
[2009/04/07 13:49] Zero Linden: that all said
Preview Grid (Aditi) (Starts in 10 minutes)
[2009/04/07 13:50] Wyn Galbraith: What we look like is just a pretty wrapping, communicating is part of the basic engine ot getting there
[2009/04/07 13:50] Zha Ewry: In the sense that your browser says "Hey, here's a stream of bytes, I don't know how to show you"
[2009/04/07 13:50] Zero Linden: I actually think that the community implementing OGP is quite happy to just standardize on a base set of 3D prim/object formats -- most of which we are already coding with today
[2009/04/07 13:50] Zero Linden: and standardizing on the base skeleton is not hard either
[2009/04/07 13:50] Zero Linden: and frankly all the texture formats are already standardized
[2009/04/07 13:50] Zero Linden: As for chat and IM -
[2009/04/07 13:50] Wyn Galbraith: pretty much Zero
[2009/04/07 13:50] Zero Linden: those would be the very first things in OGP
[2009/04/07 13:50] Zha Ewry: (well, I'm just as happy to standardize on the content mark up, to the poitn where we can do hand off, to other clients too, but exaclyy, Zero)
[2009/04/07 13:50] Zero Linden: after teleport and authentication
[2009/04/07 13:51] Annie Obscure: aorry folks, RL call
[2009/04/07 13:51] Annie Obscure: calls
[2009/04/07 13:51] Wyn Galbraith: I'd be happy to be a cube anywhere for the first time, as long as it works
[2009/04/07 13:51] Larry Massey: if these are solved problems, then writing them up as standards and publishing them should be easy deliverables in your IETF working group charter
[2009/04/07 13:51] Annie Obscure: see ya
[2009/04/07 13:51] Wyn Galbraith: B-bye!!! :D
[2009/04/07 13:52] Zero Linden: Larry - indeed - I think they would be
[2009/04/07 13:53] Zero Linden: I think the main values of bringing this to the IETF is that it will get this community to write them down in a formalized way
[2009/04/07 13:53] Morgaine Dinova: Zero: in case it's got lost in the tangential topic, here's a recap: AD of world A provides an av A1 with identity. That av goes to world B (there's an A-B interop agreement), and then that av goes to world C (there's a B-C interop agreement). The question is, how is the AD of world A handling it, given that it doesn't know (initially at least) about world C.
[2009/04/07 13:53] Zha Ewry: I'd argue solved, in the sense of "yeah, its coded, b ut, boy, one could benefit from a cleaning and documenting pass"
[2009/04/07 13:54] Larry Massey: Morgaine, you're worried about proxy authentication?
[2009/04/07 13:54] Morgaine Dinova: I'm not "worried" about anything. I wantto know how OGP proposes to handle it :-)
Preview Grid (Aditi) (Starts in 5 minutes)
[2009/04/07 13:55] Larry Massey: sorry for speaking metaphorically
[2009/04/07 13:55] Mystical Demina: if the teleport process is between the client and the server, then when the client wishes to go to work C it initiates a teleport, provides endpoints and tokens that allow world C to access any assets that the clent is autorized to use. Then the worlds don't have to have agreements in place, only agreed standards
[2009/04/07 13:55] Morgaine Dinova: We haven't had this kind of topic discussed, AT ALL.
[2009/04/07 13:55] Zha Ewry: I think thinkgni about it as proxy, is unlikely
[2009/04/07 13:55] Moon Metty: / i have another meeting, see you later everyone :)
[2009/04/07 13:55] Zero Linden: slide 20
[2009/04/07 13:55] Larry Massey: well, B accepts A's identities, and C accepts B
[2009/04/07 13:56] Gypsy Paz: me too, thanks Zero and everyone ;)
[2009/04/07 13:56] Larry Massey: so you *could* do the handoff through some kind of transitive agreement
[2009/04/07 13:56] Techwolf Lupindo: I'me going to the beta grid meeting next.
[2009/04/07 13:57] Zero Linden: So - B is a red herring here
[2009/04/07 13:57] Zero Linden: and has nothign to do with the example
[2009/04/07 13:57] Twisted Laws: are there any draft rfc's written?
[2009/04/07 13:57] Morgaine Dinova: Mystical: yeah, but the AD in world A provides the identity for the av, and that identity is used in world C too
[2009/04/07 13:57] Zero Linden: where you are coming from is not relevant in the OGP TP protocol
[2009/04/07 13:58] Larry Massey: i don't understand the role of OGP in something like cross-domain authorization
[2009/04/07 13:58] Zha Ewry: (well, especially, since you got there, so you can talk to tthere)
[2009/04/07 13:58] Zero Linden: so - user U has told their Agent domain , A, to teleport them to some url that points toa region in region domain C
[2009/04/07 13:58] Larry Massey: OAuth already is working on cross-domain authentication, why not use it, instead of inventing something new just for virtual worlds?
[2009/04/07 13:58] Zero Linden: A has no knowledge of C, and no existing trust relationship
[2009/04/07 13:58] Zero Linden: while connecting, (A contacting C), A may receive certs from C that prove C's dientity
[2009/04/07 13:59] Zero Linden: and, posisbly, offer voucing parties that may vouche for trust in C
[2009/04/07 13:59] Morgaine Dinova: Zero: It has to be relevant,since that's the most basic interop scenario :-) You have to handle the issue of one of your identities moving outside of your known space, and still being usable :-)
[2009/04/07 13:59] Larry Massey: i think that's the kind of push-back you'll get from others in the IETF as well. a single protocol mechanism *just* for cross-world interoperability, rather than using existing cross-domain mechanisms
[2009/04/07 13:59] Larry Massey: why not let me log into SL with my facebook account
Preview Grid (Aditi) (Starts now)
[2009/04/07 14:00] Zero Linden: morgaine - no - that is nonsenscial
[2009/04/07 14:00] Zero Linden: so -
[2009/04/07 14:00] Zero Linden: first things first
[2009/04/07 14:00] Zero Linden: in OGP
[2009/04/07 14:00] Zero Linden: there isn't a concept of "virtual world"
[2009/04/07 14:00] Zan Fang: can some one help me get out of here
[2009/04/07 14:00] Morgaine Dinova: Zero: unless you explain why, just saying it's nonsensical doesn't make it so :-)
[2009/04/07 14:00] Zero Linden: because it doesn't tie avatars to 3D spaces
[2009/04/07 14:00] Saijanai Kuhnri: Ashton: teleport to home
[2009/04/07 14:00] Saijanai Kuhnri: Ashton: usually has to work
[2009/04/07 14:00] Saijanai Kuhnri: Ashton: else, logout and come back on home possition
[2009/04/07 14:00] Zero Linden: In OGP
[2009/04/07 14:01] Zero Linden: there are domains for an avatar
[2009/04/07 14:01] Zero Linden: which is like a user's account
[2009/04/07 14:01] Zero Linden: so - if Facebook wants to implement that OGP Agent domain protocols
[2009/04/07 14:01] Zero Linden: then, Larry, ALL Facebook acounts can teleport to any OGP region
[2009/04/07 14:02] Larry Massey: why put OGP in the critical path?
[2009/04/07 14:02] Zero Linden: in otherwords - implementing the OGP Agent domain side is what it takes to make any concept of identity or account an entity (we say 'agent') tht can visit regions in the grid
[2009/04/07 14:03] Zha Ewry: Roughlu, because, you don't "go places" in the web
[2009/04/07 14:03] Zha Ewry: you "authenticate to get a chunk of data"
[2009/04/07 14:03] Zero Linden: because an agent in the metaverse requires more than just an account name and authentication credentials
[2009/04/07 14:03] Zha Ewry: answers in parallel
[2009/04/07 14:03] Larry Massey: and an avatar?
[2009/04/07 14:03] Zero Linden: in particular, it requires presence,
[2009/04/07 14:03] Zha Ewry: (Actually, the avatar, is something of a side effect, tho a relevent one)
[2009/04/07 14:03] Larry Massey: well, facebook has chat and presence, actually
[2009/04/07 14:04] Zha Ewry: Only sort of
[2009/04/07 14:04] Zero Linden: 'avatar' is just a representation
[2009/04/07 14:04] Larry Massey: and it has my face, which is good enough for an avatar
[2009/04/07 14:04] Zero Linden: Yes, Larry, it does
[2009/04/07 14:04] Zha Ewry: it doesn't say "Cha ewry is looking at Joe's current set of pictures"
[2009/04/07 14:04] Zero Linden: but it doesn't have "location" presence
[2009/04/07 14:04] Zero Linden: your facebook account can't visit a place
[2009/04/07 14:04] Zero Linden: which is what is different, at the core, about an agent
[2009/04/07 14:04] Zha Ewry: And it's limited to only knowing when I last fetched a URL
[2009/04/07 14:04] Larry Massey: is visiting Zero Linden's office hours
[2009/04/07 14:04] Zero Linden: from an account
Preview Grid (Aditi) (Started 5 minutes ago)
[2009/04/07 14:05] Zero Linden: yes - but your facebook account isn't
[2009/04/07 14:05] Zha Ewry: More importanrtly, I can find Larry, at Grasmere 175,100,26
[2009/04/07 14:05] Zero Linden: even if you are logged in
[2009/04/07 14:05] Zha Ewry: and I can route an IM to your agent via
[2009/04/07 14:05] Zha Ewry: ou are at 262575.6, 257903.0, 26.8 in Grasmere located at sim5665.agni.lindenlab.com (8.2.34.221:12035)
[2009/04/07 14:05] Larry Massey: i'm playing this against LambdaMOO, which has locations and avatars but no graphics
[2009/04/07 14:05] Zha Ewry: which is where things get interesting
[2009/04/07 14:06] Zero Linden: So - it *IF* Facebook wants to add the features that their accoutns can be full-fledged agents - that is, allow them to visit locations in the metaverse (hate that term)
[2009/04/07 14:06] Zero Linden: and have location presence
[2009/04/07 14:06] Larry Massey: so there's a minimum you have to implement?
[2009/04/07 14:06] Zero Linden: they'd add OGP to their mix of protocols
[2009/04/07 14:06] Zha Ewry: Well, also have "live" endpoints
[2009/04/07 14:06] Morgaine Dinova: patiently waits for this second tangent from Larry to end ... :-)
[2009/04/07 14:06] Zero Linden: because - well, the current protocols don't do this sort of presence at all
[2009/04/07 14:06] Zha Ewry: Its not exactly a tangent, Morgaine
[2009/04/07 14:06] Saijanai Kuhnri: Ashton: i think its saver not to pass LL's Ips, since it just could be possibel that a griefer gets that info somehow and misuses the info
[2009/04/07 14:07] Zha Ewry: Well, you don't pass them naked
[2009/04/07 14:07] Larry Massey: wait, XMPP is all about presence and privacy
[2009/04/07 14:08] Zero Linden: but not this sort of presence, at all
[2009/04/07 14:08] Zero Linden: it is about the presence of your client
[2009/04/07 14:08] Morgaine Dinova: Well like Zero said, in virtual worlds we "Go Places", so Facebook's relevant to this is a bit tangent as far as TP goes. It's only relevant is as a source of identity. Now that's fine, Facebook could implement an AD providing identity. But then we just get back to my question about world hopping.
[2009/04/07 14:08] Zero Linden: your account never "leaves" your jabber server
[2009/04/07 14:08] Larry Massey: i don't think that is clear at all
[2009/04/07 14:08] Larry Massey: you care about the presence of an identity
[2009/04/07 14:09] Zero Linden: presence in XMPP is about is the client connected or no
[2009/04/07 14:09] Larry Massey: presence has location, and location is relative to a namespace or coordinate system or other geographic vocabulary
[2009/04/07 14:09] Zero Linden: or what state is the client in ("away", "sleeping", "typing")
[2009/04/07 14:09] Larry Massey: i think that's a very narrow interpretation
[2009/04/07 14:09] Larry Massey: well, here, there is 'busy' or 'away' or 'logged off' or 'connected'
[2009/04/07 14:10] Morgaine Dinova: Let's just take it as a given, that ANYONE can become a source of identity if they run an AD ... Facebook included. They don't need to run a world, which is my exact point in this topic --- world A's identity provisions cannot prevent interop between worlds B and C.
[2009/04/07 14:10] Larry Massey: i'm not sure what the difference is, really, except that there's a SL coordinate space
[2009/04/07 14:10] Zero Linden: if you and I both add Geo lcoation coords in our XMPP presence - it isn't actually saying anything about us being in the same location - or sharing the same expererience
[2009/04/07 14:10] Larry Massey: yes, like that Morgain
[2009/04/07 14:10] Larry Massey: oh wait, i was thinking about geopriv
[2009/04/07 14:11] Zero Linden: whereas - when we are both placed within a region, in OGP - we are placed within the same server - sharing the same event stream
[2009/04/07 14:11] Larry Massey: sorry, i said XMPP
[2009/04/07 14:11] Larry Massey: in WoW, there are shared worlds and 'instances'
[2009/04/07 14:11] Saijanai Kuhnri: Ashton: wouldn't it be handy to add a special 'meeting' option? so that when yuo have a meeting or soemthing simular, only the direct normal chat range can talk to you and you will not get any IM's or 'shout' messages, might be a good thing to have for alot of us
[2009/04/07 14:11] Larry Massey: systems like webex and connect also have 'rooms'
[2009/04/07 14:12] Mystical Demina: i don't see what the process it took me to log into B, would be any difference than aht process to now move into C as far as a teleport, that is the inly scaable model
[2009/04/07 14:12] Larry Massey: as did placeware
[2009/04/07 14:12] Zero Linden: any way - Morgain- given the OGP protocol - where does B come into tplay during the teleport of a agent from A into a region in C
[2009/04/07 14:12] Zero Linden: ?
[2009/04/07 14:12] Morgaine Dinova: Zero: you said "OGP doesn't deal in 'worlds'", but OGP will have to do that in MMOX, because everyone thinks in terms of "worlds" for VW interop, not grids and regions.
[2009/04/07 14:12] Larry Massey: my battery is dead and i didn't bring my charter with me, so i'm going o sign out :P
[2009/04/07 14:12] Zero Linden: well - we'll have to do a better job of explaining then
[2009/04/07 14:12] Larry Massey: happy to talk about this more, either in-world or off
[2009/04/07 14:12] Zero Linden: later Larry
[2009/04/07 14:13] Morgaine Dinova: Zero: B comes into play by having an interop agreement with C. World A doesn't have one.
[2009/04/07 14:13] Zero Linden: what does an interop agreement ebtween two region domains mean, here?
[2009/04/07 14:13] Zero Linden: regions don't visit each other
[2009/04/07 14:13] Zha Ewry: That's policy vs. mechanism gettign entnagled
[2009/04/07 14:13] Zha Ewry: The mechanism doesn't change
[2009/04/07 14:14] Saijanai Kuhnri: Ashton: the 'meeting' option what i mean, should be something between 'normal' online mode and 'busy', think it is atleast handy for those that need concentration for building/scripting. Ofcourse the name can be changed in this!
[2009/04/07 14:14] Zha Ewry: the region may reject a connectoin based on policy
[2009/04/07 14:14] Morgaine Dinova: Zero: not the regions, the av. The av with identity from A travels to B and to C
[2009/04/07 14:14] Zero Linden: and if you mean C presents a cert that says "B vouches for me" - then A checks to see if it's trust of B includes vouching
[2009/04/07 14:14] Zero Linden: but again - this is just about how A decides what set of abilities it will allow C
[2009/04/07 14:14] Mystical Demina: if that is true morgain, than B has to act as a proxy to C
[2009/04/07 14:14] Zha Ewry: and wthere it will open a connection to C at all
[2009/04/07 14:15] Zero Linden: yes - so the agent domain of B (if any) and of C (if anY) doesn't come into play at all
[2009/04/07 14:15] Zha Ewry: Which again is policy, not mechanism
[2009/04/07 14:15] Zha Ewry: is Away from keys a moment
[2009/04/07 14:15] Morgaine Dinova: Zero: A doesn't know about C at all (at least initially), but it's providing identity, and that identity will be used (because of B-C) agreement in world C
[2009/04/07 14:15] Morgaine Dinova: Ignore policy, we're talking about mechanisms.
[2009/04/07 14:16] Zero Linden: x509 from C, signed by B, with B vouching for C's trustworthyness
[2009/04/07 14:17] Zero Linden: x509 from B, vouching for it's own identity
[2009/04/07 14:17] Morgaine Dinova: There are several options, I'm just asking which option OGP expresses. One option is that A denies the B-C interop, ie. does not allow the identity to port. I'm hoping that that's not the case, or else OGP is dead. :-)
[2009/04/07 14:17] Zero Linden: pre-arranged agreement between B and A where A agrees to trust who B vouches for
[2009/04/07 14:17] Zero Linden: done
[2009/04/07 14:17] Zero Linden: B doesn't even need to be a region or agent domain
[2009/04/07 14:17] Zero Linden: B is Verisign!
[2009/04/07 14:18] Zero Linden: or some such
[2009/04/07 14:18] Larry Massey: battery hasn't died yet. but why not use OAuth for authentication?
[2009/04/07 14:18] Zero Linden: is that clear mechanism
[2009/04/07 14:18] Morgaine Dinova: No, B is just a world with different arrangements to A or to C, but nevertheless running OGP
[2009/04/07 14:18] Morgaine Dinova: It's not Verisign
[2009/04/07 14:18] Zha Ewry: 'The mechanism, is the same in all cases
[2009/04/07 14:18] Zha Ewry: the policy is going to be A's
[2009/04/07 14:18] Zero Linden: Larry - we could possibly - but the Oauth model might be too bound up in the web, and assuming a user agent (save the 2 legged case)
[2009/04/07 14:19] Zha Ewry: does it wanr to talk to C
[2009/04/07 14:19] Zero Linden: and OAuth wasn't around when we built all this
[2009/04/07 14:19] Zha Ewry: What is its policy towards a regoin is doesn't know about?
[2009/04/07 14:19] Zha Ewry: (Which probably means, what is its policy to a region, that doesn' present any proof of trust)
[2009/04/07 14:19] Zha Ewry: Regoin C, has a URL
[2009/04/07 14:19] Zha Ewry: you do an get seed cap on it
[2009/04/07 14:19] Zero Linden: Morgain - there is no OGP involved in the determination of A's trust of C, other than fields in X509
[2009/04/07 14:20] Zha Ewry: (or you deciced you don't like any of its X.509 certs)
[2009/04/07 14:20] Zha Ewry: at which point, you get to decide if you want to talk to it via http, I suppose
[2009/04/07 14:20] Morgaine Dinova: Can I take this line "[14:17] Zero Linden: pre-arranged agreement between B and A where A agrees to trust who B vouches for" to mean that the AD of A will pass the identity transitively anywhere, without regard to direct trust agreements? (REMEMBER that C could be Z .... a huge chain of links away)
[2009/04/07 14:20] Zero Linden: Also, OAuth would still requires us to define a full set of token ids, and methods....)
[2009/04/07 14:21] Zha Ewry: That isn't a poliicy I'm setting in *my* AD, but you could
[2009/04/07 14:21] Zero Linden: we use capabilities for most of what OAuth offers
[2009/04/07 14:21] Morgaine Dinova: Remember that world B cannot possibly vouch for world Z, far down theline
[2009/04/07 14:21] Zero Linden: (which, if there is no browser involves, somewhat cleaner)
[2009/04/07 14:21] Zha Ewry: The only interesting questoin, in many ways, is "What policy does A's AD have to C"
[2009/04/07 14:21] Zero Linden: well - look at your browser
[2009/04/07 14:21] Zero Linden: how long of chain does it trust?
[2009/04/07 14:21] Zero Linden: usually 3 or 4 at most --
[2009/04/07 14:22] Morgaine Dinova: Browsers don'r require agreements between the various sites you visit :P
[2009/04/07 14:22] Zero Linden: and usually the top chain of 2 or 3 are all the same company or affliats with agreements
[2009/04/07 14:22] Zero Linden: actually
[2009/04/07 14:22] Zero Linden: it is the same as the question of "You are on site B, and you follow a link to site C, which your browser has no cert for nor has never heard of before"
[2009/04/07 14:22] Morgaine Dinova: Zha: that's well put. But remember that C could be Z.
[2009/04/07 14:23] Zero Linden: The big problem here, is that there isn't a written up proposal or case to refute
[2009/04/07 14:24] Zero Linden: I think the answer the question is "x509 certs and policy"
[2009/04/07 14:24] Zha Ewry: In general
[2009/04/07 14:24] Zha Ewry: I think its a matter of exactly that
[2009/04/07 14:24] Morgaine Dinova: Zero: bad analogy, because there are very few cert authorities used on the web, and using new ones is hugely broken. Whereas there will be zillions of identity providers and worlds.
[2009/04/07 14:24] Mystical Demina: to me if client logs into A, and A supports the client to teleport off the grid into a new grid than it will provide the client with any needed tokens to allow who the client provides them to access to the servers of A that the client has subscribed into. So when client teleports into B if B and A have a trust I will get authenticated into B, if A and B don not have a trust then I must provide credentials to get into B. So authnetication to enter a space is one thing. Second part is client providing URL to its avatar servers. So if client wishes to go to C it will need to get authenticated again and could provide the token from B, or have to provide credentials. Once authneticated the client provides URLS and tokens to C for its identity which A be definiotn is provider as a service for the cient.
[2009/04/07 14:25] Zha Ewry: I think that we'll have policies which are all over the map
[2009/04/07 14:25] Zero Linden: Morgain - I don't know that I agree with that
[2009/04/07 14:25] Zha Ewry: The protocol, is easy, the X.509 stuff is easy, the policies will be hard
[2009/04/07 14:25] Zero Linden: Mystical
[2009/04/07 14:25] Zero Linden: no
[2009/04/07 14:25] Zero Linden: that is exactly what we are trying to avoid totally
[2009/04/07 14:25] Mystical Demina: but i see this will not work in the short term
[2009/04/07 14:25] Zero Linden: we don't want people to have to have accounts on each region domain
[2009/04/07 14:26] Mystical Demina: and is not scalabel any other way
[2009/04/07 14:26] Zero Linden: and perhaps this is the crux of Morgain's confusion
[2009/04/07 14:26] Mystical Demina: the world is going browers based, and cant' require everyone in the worldto have trust
[2009/04/07 14:26] Zero Linden: there is no account for the Agent from A on B or C
[2009/04/07 14:26] Zero Linden: none
[2009/04/07 14:26] Wyn Galbraith: Need a federation
[2009/04/07 14:26] Morgaine Dinova: Zero: you'll be glad to hear that we've smoothly migrated to my 2nd topic now :PPPP
[2009/04/07 14:27] Zero Linden: well then
[2009/04/07 14:27] Zero Linden: the second topic is too nebulous to discuss
[2009/04/07 14:27] Mystical Demina: if I log iot A and A is my avatar provider serers, than I will ned to proviuder those endpoints any where i go
[2009/04/07 14:27] Morgaine Dinova: Zero: Infinity areed with it
[2009/04/07 14:27] Morgaine Dinova: [3]
[2009/04/07 14:27] Zero Linden: This isn't the right forum for discussing what is a point of view
[2009/04/07 14:27] Morgaine Dinova: So if it's too nebulous for you, perhaps you need to read the many posts about it :-)
[2009/04/07 14:28] Zero Linden: without there being some clear porotocol or proposal to discuss
[2009/04/07 14:28] Zero Linden: do you have an alterantive to the trust model?
[2009/04/07 14:28] Morgaine Dinova: In a Groupies meeting, Zha said it would have to be addressed with nil trust as an option.
[2009/04/07 14:29] Zero Linden: I think nil is a fine trust option
[2009/04/07 14:29] Zha Ewry: I expect there will be lots of "nil"
[2009/04/07 14:29] Zha Ewry: and that lots of services will respond with alt content, or 401s
[2009/04/07 14:29] Zero Linden: I visit the vast majority of my web sites with nil as a trust option
[2009/04/07 14:29] Morgaine Dinova: Like Infinity agreed (I even got a +! for it :P), trust becomes meaningless when scaled to millions. All you are doing then is empty paperwork.
[2009/04/07 14:30] Zha Ewry: Again, the protocol seperate the two out, and we can set them out as policy
[2009/04/07 14:30] Zha Ewry: Nobody is suggesting that we do deep trust across milllions
[2009/04/07 14:30] Zero Linden: I don't think tansitive trust will be extended very deep
[2009/04/07 14:30] Zha Ewry: but., the speeration lets us do deep trust, when we need it
[2009/04/07 14:30] Zero Linden: really -
[2009/04/07 14:30] Morgaine Dinova: OK, I sense progress then.
[2009/04/07 14:31] Zero Linden: I know how to get teleport to work - and I know how to separate out trust - and I know how to do trust with x509 certs in a few clear models
[2009/04/07 14:31] Zero Linden: we have made them be pretty decoulpled
[2009/04/07 14:31] Zero Linden: so that things can use new trust frameworks as they come along
[2009/04/07 14:31] Zero Linden: but I haven't heard anyone actually proposeing one
[2009/04/07 14:31] Morgaine Dinova: What we're saying is that (i) trust agreements will only have short transitive behaviour. (ii) The identity does not have restrictions on movement only to worlds with a direct trust link.
[2009/04/07 14:32] Zero Linden: these ahve always been so
[2009/04/07 14:32] Morgaine Dinova: Does that sound right?
[2009/04/07 14:32] Saijanai Kuhnri: Ashton: dont ask me, only thing i know you all talking about C, C++ and simular languages :p
[2009/04/07 14:32] Zero Linden: though (i) is a matter of what the industry does, not by fiat
[2009/04/07 14:33] Zero Linden: (ii) is baked into OGP
[2009/04/07 14:33] Morgaine Dinova: Zero: well since you're never on MMOX, and only respond there when pressured, we have a lot of trouble discussing anything concrete on MMOX at all about OGP. This is precisely what I was saying about lack of contect.
[2009/04/07 14:33] Zero Linden: though any domain is free to have a policy that disallows it
[2009/04/07 14:33] Zha Ewry: well, we're not saying antying about trust lengh in the spec
[2009/04/07 14:33] Zha Ewry: We're saying its ap olicy chioce
[2009/04/07 14:33] Zha Ewry: which is as it should be
[2009/04/07 14:33] Zha Ewry: people should be allowed to do dumb tings ;_
[2009/04/07 14:33] Zha Ewry: :-)
[2009/04/07 14:34] Saijanai Kuhnri: Ashton: Zha: well they can for all i care, as long as it doesnt lagg me up or even keep me off SL :p
[2009/04/07 14:34] Morgaine Dinova: Sure, but there must be very concrete mention of transitivity of trust agreements in OGP, or else it leads to the whole scheme being questioned as unworkable.
[2009/04/07 14:34] Zha Ewry: The statements are in the policy and the X.509 rules you set
[2009/04/07 14:34] Morgaine Dinova: This needs to be tied doww, as a mechanism.
[2009/04/07 14:34] Zha Ewry: OGP gives not a whit
[2009/04/07 14:35] Saijanai Kuhn: seems to me that one simple model of trust transivity is a AD providing authorization to a hypergrid, which then takes responsibilty for trust domain issues until the avatar leaves the domain of hte hypergrid. In which case, the AD comes back into play
[2009/04/07 14:35] Zero Linden: Morgaine please don't throw out ultimatums on the requirements of OGP
[2009/04/07 14:35] Zero Linden: instead - propose an actual design
[2009/04/07 14:35] Morgaine Dinova: Zero: no ultimatums at all, just trying to engage you in technical discussion, and it's very hard work :-))))
[2009/04/07 14:36] Morgaine Dinova: Zero: OGP is YOUR baby --- tell us how it works :-)
[2009/04/07 14:36] Zero Linden: I did - x509 certs
[2009/04/07 14:36] Zero Linden: done
[2009/04/07 14:36] Zero Linden: !
[2009/04/07 14:36] Zero Linden: now - in the absence of any alternative proposal or concrete model to discuss
[2009/04/07 14:36] Zha Ewry: rolls her eyes
[2009/04/07 14:36] Zero Linden: I propose we move on
[2009/04/07 14:36] Saijanai Kuhnri: Ashton: lol, doesnt tell a noob like me much, but i think ill need to learn the scripting languages first for that :p
[2009/04/07 14:37] Morgaine Dinova: Zero: I think you'll find that saying one world will not be greeted as sufficient discussion on the topic in an IETF workgroup. :-)
[2009/04/07 14:37] Morgaine Dinova: s/world/word/
[2009/04/07 14:37] Morgaine Dinova: I get a very strong sense of wanting to brush things under the carpet, and it's not wholesome for engineering.
[2009/04/07 14:37] Zero Linden: In the ZOH of 3rd March, you asked "What else should we add to OGP?". .. Therefore I suggest "Let's add OGP: Object Protocol".
[2009/04/07 14:38] Dahlia Trimble: bye all :)
[2009/04/07 14:38] Zero Linden: I don't think anything was brushed under any carpet -
[2009/04/07 14:38] Morgaine Dinova: Ah yes, that followed from Mystical's topic, which many others have mentioned too.
[2009/04/07 14:38] Morgaine Dinova: Object transport models.
[2009/04/07 14:38] Morgaine Dinova: Decoupling of asset providion from region provision from identity provision.
[2009/04/07 14:39] Zero Linden: - you don't have a clear issue or proposal or problem or concrete model --- whereas x509 and vouching for trust via signatures on certs is quite quite well understood and deployed
[2009/04/07 14:39] Saijanai Kuhnri: Ashton: i dont have a feeling of 'brush things under the carpet' at all, actually i see improvements in several things already. that doesnt mean any other specific thing is on hold or cancelled
[2009/04/07 14:39] Saijanai Kuhnri: Ashton: atleast... thats my feeling in this
[2009/04/07 14:39] Morgaine Dinova: Zero: let's discuss it on MMOX please, in depth.
[2009/04/07 14:39] Zero Linden: I don't have any idea what "provision" means in that sentence--- povisioning as in setting account values?
[2009/04/07 14:40] Morgaine Dinova: Zero: well for example Opensim/realXtend are looking at Cable Beach for asset storage service --- decoupled asset provision.
[2009/04/07 14:40] Zero Linden: and - okay, so does anyone have a concrete written draft for object protocol? (though I'm not sure what that refers to)
[2009/04/07 14:40] Saijanai Kuhnri: Ashton: if you remember 2007/2008, then you know that SL improved alot and is much more stable, this means improvements in Zero's work also improved. We are all human, mistakes are to be learned from. that counts for everyone.
[2009/04/07 14:41] Saijanai Kuhnri: Ashton: unless you area robot, then the fool that programmed you is the fault :p
[2009/04/07 14:41] Saijanai Kuhnri: Ashton: no offence tho :p
[2009/04/07 14:41] Morgaine Dinova: The question for us here is how OGP intends to handle decoupled storage service.
[2009/04/07 14:42] Morgaine Dinova: Opensim already allows assets to be stored on S3 for example.
[2009/04/07 14:42] Zero Linden: yes - assets will be URLs - done
[2009/04/07 14:42] Zero Linden: I think we said that already at some point :-)
[2009/04/07 14:42] Saijanai Kuhnri: Ashton: i keep hearing about opensim, but i still dont know what it is.
[2009/04/07 14:42] Zero Linden: I don't think OGP intends to define storage beyond that
[2009/04/07 14:42] Zero Linden: BUT
[2009/04/07 14:43] Zero Linden: I don't have drafts yet for that
[2009/04/07 14:43] Zero Linden: Is cable beach Adam's multi-mysql-store thing?
[2009/04/07 14:43] Zha Ewry: No, Cable is intels
[2009/04/07 14:44] Zero Linden: oh
[2009/04/07 14:44] Zero Linden: well - is there a proposal?
[2009/04/07 14:44] Wyn Galbraith: Interesting
[2009/04/07 14:44] Morgaine Dinova: Sure, we're not asking for ready drafts, we're very reasonable :-) ... we're simply asking for discussion about it on MMOX.
[2009/04/07 14:44] Zha Ewry: The place where it may get tricky, is inventory vs. asset
[2009/04/07 14:44] Zero Linden: would that be visible at the OGP level?
[2009/04/07 14:44] Zero Linden: asset = pile o' immuable bits
[2009/04/07 14:44] Zero Linden: inventory = protected reference with permissions to asset = asset capability!
[2009/04/07 14:45] Zero Linden: well, asset capability w/metadata
[2009/04/07 14:45] Zero Linden: so - issue is?
[2009/04/07 14:45] Morgaine Dinova: Ready solutions imposed from LL is what is causing all the trouble in MMOX --- the total absense of discussion about the OGP model is a severe problem there.
[2009/04/07 14:46] Morgaine Dinova: Zero: why not join us and discuss solutions? Or the *problem*, even?
[2009/04/07 14:46] Zha Ewry: Not so clear that works, entirely cleanly
[2009/04/07 14:46] Zero Linden: I'm here to discuss OGP
[2009/04/07 14:47] Morgaine Dinova: We all are. :-)
[2009/04/07 14:47] Zero Linden: I don't know what the "ready solutions imposed from LL" is about - but it is clearly a dig
[2009/04/07 14:47] Zha Ewry: Not clear to me that the "permissions" seperated out is entirely what you want in web scale
[2009/04/07 14:47] Zero Linden: what are you trying to say? URLs to HTTP chuncks is an "LL ready solution"?
[2009/04/07 14:47] Zha Ewry: But.. a proposal and some discussion is the way forward with that
[2009/04/07 14:47] Morgaine Dinova: Zero: it's not a dig, you replied that "you don't have a draft ready" ... well I know, it's too early for drafts. We just want discussions about what's intended to be the design.
[2009/04/07 14:48] Morgaine Dinova: If no discussion, and then a ready solution is possed out, that's "ready soluitions imposed by LL".
[2009/04/07 14:48] Morgaine Dinova: s/possed/popped/
[2009/04/07 14:49] Zero Linden: well - in *this* forum, the plan was talk about substantive work being done in defining more of OGP -
[2009/04/07 14:49] Morgaine Dinova: The rules in MMOX are differnet to the rules in AWG. In AWG you provided "fait accompli", but the IETF isn't like that. Technical discussion is required about designs.
[2009/04/07 14:50] Morgaine Dinova: Cool! Hence my Topic #3!
[2009/04/07 14:50] Morgaine Dinova: OGP: Object Protocol
[2009/04/07 14:50] Zero Linden: it is not intended as an open ended conversation -
[2009/04/07 14:50] Zero Linden: I don't have anything to contribute on that at this time
[2009/04/07 14:51] Morgaine Dinova: Noted.
[2009/04/07 14:51] Zero Linden: Nor does anyone else - what do you want to say about it?
[2009/04/07 14:51] Zero Linden: And when I suggest where we have talked about the genreal form of such things (as we did in 2007 and 2008 -- we touched on ALL of this)
[2009/04/07 14:51] Zero Linden: you tell me "ready made LL designs"
[2009/04/07 14:51] Zero Linden: which is unfair
[2009/04/07 14:51] Morgaine Dinova: Zero: I suggest that you join us in MMOX discussion on the topic then, so that you have material for your drafts that can meet interop requirements.
[2009/04/07 14:52] Morgaine Dinova: I readily accept that we don't have the material ready yet.
[2009/04/07 14:52] Morgaine Dinova: But lots of people on MMOX and in Opensim *are* talking about object interop, so there's a ready source of ideas and requirements there.
[2009/04/07 14:53] Morgaine Dinova: And it's very good material.
[2009/04/07 14:53] Morgaine Dinova: If you like, we can end my 3 points there.
[2009/04/07 14:53] Zero Linden: good
[2009/04/07 14:53] Zero Linden: timing
[2009/04/07 14:53] Twisted Laws: use the client/user for storage of items that can cross worlds and are not rez'd in a world
[2009/04/07 14:53] Zero Linden: it is
[2009/04/07 14:53] Zero Linden: 3pm
[2009/04/07 14:53] Morgaine Dinova: Twisted: yep, many people have suggested that in MMOX
[2009/04/07 14:54] Zero Linden: thank you all for coming
[2009/04/07 14:54] Wyn Galbraith: Thanks Zero always educational :)
[2009/04/07 14:54] Twisted Laws: thanks
[2009/04/07 14:54] Zha Ewry: Thanks Zero
[2009/04/07 14:54] Mystical Demina: enjoyed it, thanks for the feedback and opportunitity to share
[2009/04/07 14:54] Morgaine Dinova: 2 hours per month is not enough, we've barely started. :-(
[2009/04/07 14:54] Zero Linden: till next time
[2009/04/07 14:55] Saijanai Kuhnri: Ashton: ty for your time and the bear Zero!
[2009/04/07 14:55] Morgaine Dinova: waves to Zero

User:Zero Linden/Office Hours/2009 apr 07

Navigation menu

Search