AW Groupies/Chat Logs/AWGroupies-2009-03-03

[7:45] Connecting to: in-world Voice Chat...
[7:45] Connected undefined:
http://slurl.com/secondlife/Longfellow/198/59/25 (Starts in 10 minutes)
http://slurl.com/secondlife/Longfellow/198/59/25 (Starts in 5 minutes)
http://slurl.com/secondlife/Longfellow/198/59/25 (Starts now)
http://slurl.com/secondlife/Longfellow/198/59/25 (Started 5 minutes ago)
[8:53] Morgaine Dinova: Bah, no sound. I bet flash ate it.
[8:53] Morgaine Dinova: kills FF
[8:56] Latha Serevi: Morgaine, if you make the IRC relay "buy original for L$0" then it's turn-offable by someone else. That's what I suggest if we leave the premises with the microphone on. :-)
[8:56] Morgaine Dinova: Woohoo, sound! (/me hates the damn flash)
[8:57] Morgaine Dinova: Latha, please pick it up and put another down --- I only stuck one down because when I was testing from another region, it disabled the old one :-(
[8:57] Latha Serevi: You're the only one who can pick this one up; see my suggestion above. :-)
[8:57] Morgaine Dinova: Deleted :-)
[8:58] Latha Serevi: Check. Check. (I notice that it isn't active until someone in SL says something)
[9:00] LathaSerevi barks:
[9:01] Morgaine Dinova: I set the gateway to AW Groupies ownership earlier, assuming that that would allow people to admin it, prolly wrong.
http://slurl.com/secondlife/ThorneBridgeTown/155/129/23 (Starts in 15 minutes)
[9:17] sensory Hax: greetings all
[9:17] [[User:xsensory Hax[it>en]|xsensory Hax[it>en]]]: greetings all
[9:17] SharedRealm Engineer: morning
[9:18] panelope Vella: good morninig
[9:18] LathaSerevi howls: from IRC
http://slurl.com/secondlife/ThorneBridgeTown/155/129/23 (Starts in 10 minutes)
http://slurl.com/secondlife/ThorneBridgeTown/155/129/23 (Starts in 5 minutes)
http://slurl.com/secondlife/ThorneBridgeTown/155/129/23 (Starts now)
[9:33] Morgaine Dinova: Hiyas all :-)
[9:33] SharedRealm Engineer: hi
[9:33] Teravus Ousley: Hello
[9:33] sensory Hax: Hello sir
[9:33] [[User:xsensory Hax[it>en]|xsensory Hax[it>en]]]: Hello sir
[9:34] Morgaine Dinova: Hehe, Hax is on both
[9:34] Morgaine Dinova: /names
http://slurl.com/secondlife/ThorneBridgeTown/155/129/23 (Started 5 minutes ago)
[9:35] sensory Hax: ????
[9:35] [[User:xsensory Hax[it>en]|xsensory Hax[it>en]]]: ?
[9:35] Morgaine Dinova: Hi Dahlia
[9:36] Dahlia Trimble: Hi :)
[9:36] sensory Hax: hello
[9:36] [[User:xsensory Hax[it>en]|xsensory Hax[it>en]]]: hello
[9:36] Teravus Ousley: Hello Dahlia
[9:36] Morgaine Dinova: Hiya Zha
[9:37] Zha Ewry: Morning all
[9:37] Zha Ewry: Sorry I'm running late, you can blame the blue Kitty
[9:37] LathaSerevi: I: suggest that we submit to the iron fist of Fearless Leader (Zha) in order to stay focused today.
[9:37] Morgaine Dinova: It's OK, this kittie has blue armour underneath ;-)
[9:37] Zha Ewry: looks in her inventory for "Fist, Iron"
[9:37] Teravus Ousley: Hey Zha, I detailed a few technical nit picky items that need to be fixed on the draft.. and, suggested that using the word eXperience might dictate that certain statements may need to be changed slightly.
[9:38] Rex Cronon: hello everyboyd
[9:38] Morgaine Dinova: Hiya Rex
[9:38] Teravus Ousley: overall, great draft though
[9:38] sensory Hax: hello rex
[9:38] [[User:xsensory Hax[it>en]|xsensory Hax[it>en]]]: hello rex
[9:39] Rex Cronon: hi:)
[9:39] Zha Ewry: Saw that, and all good gets
[9:39] Zha Ewry: Beating the thing into IETF submissino format was ummm. Tedious
[9:39] Teravus Ousley: haha, I bet
[9:39] Teravus Ousley: good result though
[9:40] Morgaine Dinova: Heh. Used nroff -man? ;-))))
[9:40] Zha Ewry: I wish
[9:40] Zha Ewry: I re-entered it right into XML
[9:40] Zha Ewry: <list style="numbers">
[9:40] Zha Ewry: <t>
[9:40] LathaSerevi: Just: so everyone's on the same page -- topic 1 for the day is likely to be Zha's screed as internet-draft: [1]
[9:40] Zha Ewry: This is not the best way
[9:40] Morgaine Dinova: Ouch
[9:40] Zha Ewry: </t>
[9:40] LathaSerevi: .txt:
[9:40] Zha Ewry: <t?
[9:41] Zha Ewry: but it does ensure you submit a valid format
[9:41] Zha Ewry: </t>
[9:41] Zha Ewry: </list>
[9:41] Zha Ewry: <emote>
[9:41] Rex Cronon: i have seen something weird on the mmox page
[9:41] Zha Ewry: What a major pain in the backside
[9:41] Zha Ewry: </emote>
[9:41] SharedRealm Engineer: haha
[9:41] Rex Cronon: there is a list of known vw that r intereted in interop
[9:42] Rex Cronon: but, there is nobody from darkstar?
[9:42] Zha Ewry: A number are missing
[9:42] Teravus Ousley: heh heh, a normative reference for DOMINATION... *checks the draft for use of that word*
[9:42] Teravus Ousley: no... I guess that's just in there for jokes?
[9:43] Morgaine Dinova: Darkstar would be good. I guess anyone could just knock off an email to them. It's not as if there were a formal procedure to be followed :P
[9:43] Teravus Ousley: oh wait.. that's an Informative Reference
[9:43] Zha Ewry: OK, so....
[9:44] Rex Cronon: darkstar is also related to wonderland and i finding even stranger that sun is not interested:)
[9:44] Morgaine Dinova: Mind you, the big puzzling thing is why Qwaq isn't on MMOX, seeing as they were extremely interested in interop at the meeting.
[9:44] Zha Ewry: Yes, anyone who has comments about the "problem/scope" statement is welcome to raise them.. I actually had a subtly different agenda item as well
[9:44] Zha Ewry: (and Mark Cahill is on, and he is that community)
http://slurl.com/secondlife/SLVEC/190/144/22 (Starts in 15 minutes)
http://slurl.com/secondlife/Ambleside/181/99/30 (Starts in 15 minutes)
[9:45] Rex Cronon: did ll do a press releas about mmox? maybe more people would get interested
[9:46] Zha Ewry: So.. I finally sorted out in my head, what Jon Watte's on about.. and wanted to talk about that, degrees of interop, and some oddities I've noticed people assuming, or quasi assuming the past few weeks
[9:46] LathaSerevi: Okay,: Zha, let's hear it!
[9:47] Zha Ewry: And.. finally, there is this interestign tension, between the boring grunt work of doing some things which are less broad, vs the bigger space
[9:47] Zha Ewry: So..
[9:47] Dahlia Trimble: there doesn't seem to be a good definition for "interop"
[9:48] Meadhbh Oh: yes. there are several good definitions for "interop"
[9:48] Zha Ewry: Not at all, Dahlia, and it's not helpful
[9:48] Zha Ewry: Its the lack of consensus
[9:48] LathaSerevi: (I: mentioned on irc://chat.freenode.net/opensim that people could come visit with us at this meeting via irc://irc.quickfox.net/mmox ... hope I don't get killed.)
[9:48] Meadhbh Oh: the space is big though
[9:48] Meadhbh Oh: there's room for lots of consensus
[9:48] Zha Ewry: And, to an extent that we have interop in the large, small, and medium lurking
[9:48] Meadhbh Oh: we don't need one big shared view of consensus
[9:48] Zha Ewry: So.. Agenda:
[9:48] Zha Ewry: What Jon's been on about
[9:49] Rex Cronon: how about this. interop=the ability to go between VWs without major headaches:)
[9:49] Meadhbh Oh: nope
[9:49] Meadhbh Oh: i think we're going after definition 5
[9:49] Meadhbh Oh: some VWs share some protocol
[9:49] Zha Ewry: Some "odd ideas" which I want to get out on the table and talk about (such as the lurking Uber-client" idea which keeps surfacing)
[9:49] Dahlia Trimble: which one is 5 now?
[9:49] Meadhbh Oh: cause definition 1 is gonna take a LONG timne
[9:49] Zha Ewry: and.. then.. how to do some practical stuff, possibly in parallel
[9:49] Meadhbh Oh: some VWs share some protocol == def 5
http://slurl.com/secondlife/SLVEC/190/144/22 (Starts in 10 minutes)
http://slurl.com/secondlife/Ambleside/181/99/30 (Starts in 10 minutes)
[9:50] Meadhbh Oh: at the end of the day we can't force everyone to use the same protocol
[9:50] Meadhbh Oh: or protocols
[9:50] Teravus Ousley: Some would argue that the Uber Client is unrealistic.
[9:50] Zha Ewry: Nope
[9:50] Zha Ewry: And the ITEF would tell us not to try
[9:50] Rex Cronon: "uber client'?
[9:50] Zha Ewry: So..
[9:50] Meadhbh Oh: right. even in the HTTP world, we get different clients that ignore bits of teh HTTP spec
[9:50] Meadhbh Oh: and just don't offer all the services
[9:50] Zha Ewry: Jon first, then uber client, and related oddities
[9:50] Teravus Ousley: .. one reason why the Uber Client is unrealistic.. is because there hasn't been enough research into the 3D space and what sort of interactions and interfaces work well with it.
[9:51] Meadhbh Oh: *cough*IE*cough*webDAV*
[9:51] Teravus Ousley: ah.. ok. Jon first.
[9:51] SharedRealm Engineer: uber client=browser for VWs
[9:51] Meadhbh Oh: is jon here?
[9:51] Zha Ewry: So.. Jon's perspective, which is not totally off base, but.. also not entirely constructive.. is that...
[9:51] Meadhbh Oh: or is he SL-phobic enough to not want to visit?
[9:51] Zha Ewry: roughly...
[9:51] Rex Cronon: i guess the sl viewer is an "uber client". hmm:)
[9:52] Zha Ewry: "Look, all these worlds are too different to merge at the client/sim interaction level, so don't bother, we can do it all at the sim-sim level"
[9:52] Meadhbh Oh: mmm.. i don't see that's what he's on about
[9:52] Zha Ewry: Which.. is.. an interesting perspective, and has some huge problems hiding in it
[9:53] Meadhbh Oh: i think it's more.... "i like the interaction model where there's a consentual reality shared by all viewers, generated by viewers"
[9:53] Zha Ewry: Newp
[9:53] Zha Ewry: I thought that at first, but its actually not what he's pushing
[9:53] Rex Cronon: does he want a diffrent type of vw that is used to connect other types:)
[9:53] Zha Ewry: That *is* how OLIVE works
[9:53] Morgaine Dinova: The SL viewer is open source, so it'll be as open as people decide to make it. It's already the common denominator between SL and Opensim, and they're diverging, while the viewer will continue to interoperate with both.
[9:53] Meadhbh Oh: OLIVE(tm) doesn't connect to other worlds, though
[9:53] Meadhbh Oh: and he's already mentioned it will never connect to SL or OpenSom
[9:54] Meadhbh Oh: s/OpenSom/OpenSim/
[9:54] Zha Ewry: Not at the client/sim level he's not as sure at the sim-sim level.. which is sort of an interesting perspective
[9:54] Morgaine Dinova: Jon says that his users on OLIVE want interop, so that's what he's working to give them. Interop with who is unstated, but my guess is, with anyone who runs MMOX protocols.
[9:54] Zha Ewry: Let me try and quickly sumarize my understanding of how he might do it
[9:54] Meadhbh Oh: and honestly, i am VERY surprised he submitted LESS, as it sort of implies they're ceding aspects of their IPR
http://slurl.com/secondlife/SLVEC/190/144/22 (Starts in 5 minutes)
http://slurl.com/secondlife/Ambleside/181/99/30 (Starts in 5 minutes)
[9:55] Zha Ewry: (Not really, it's just the line format, I'm pretty sure the IPR they want to protect is not in there)
[9:55] Zha Ewry: So.......
[9:55] Zha Ewry: What he suggests you could do is say
[9:55] Morgaine Dinova: I'm not surprised at all. Working on MMOX requires flexibility from all participants.
[9:55] Zha Ewry: "No, we're never going to share line formats to clients, don't try"
[9:55] Meadhbh Oh: well... from the VWIF... we know we don't know what they want to disclose, so sure
[9:55] Meadhbh Oh: ROTFLMAO
[9:55] Zha Ewry: Instead, define a way for two sims to "co-simulate" a space
[9:56] Meadhbh Oh: Morgaine has obviously never worked with the forterra people
[9:56] Zha Ewry: (Jon's being more engaged here, than VWIF, but they have a reputation, yes)
[9:56] SharedRealm Engineer: He wants to use VW proxy/gateways to connect VWs
[9:56] Morgaine Dinova: Nor is it relevant whether I have worked with Forterra people. It's not us and them in MMOX.
[9:56] Zha Ewry: So, not quite gateways, but close
[9:57] Zha Ewry: and that's got all sorts of hair lurking in it
[9:57] SharedRealm Engineer: more like reverse-proxies that translate different VW objects
[9:57] Meadhbh Oh: yeah. internally i'm constantly fighting the perception that forterra is in MMOX exclusively to F*** it up
[9:57] Zha Ewry: So, he'd say, "So, if we want to share this meeting space, between OLIVE and SL, we'd create a "shared context" and one sim would pass on to the other the things which it is managing"
[9:58] Meadhbh Oh: great. sounds great. more power to them
[9:58] Zha Ewry: Which gets wicked ugly, if you want 10 sets of clients
[9:58] Zha Ewry: But...
[9:58] Zha Ewry: At the same time...
[9:58] Meadhbh Oh: i wonder if we could move towards something COLLADA like
[9:58] Zha Ewry: I will note, the set of things the two worlds would need to share is not deeply different
[9:58] Zha Ewry: from what they tend to tell thier own clients
[9:58] Meadhbh Oh: the way COLLADA is supposed to be used (according to a lot of people) is as an interop format
[9:59] Meadhbh Oh: so you don't have to represent your stuff as COLLADA, you just ahve to create a converter
[9:59] Zha Ewry: More a static interop format, but yes
[9:59] Dahlia Trimble: an interop format for a workflow
[9:59] Zha Ewry: So....
[9:59] Meadhbh Oh: maybe there's an eqivalent in the more dynamic protocol space
[9:59] Meadhbh Oh: THAT is definitely something we could work with the forterra peeps about
[9:59] Zha Ewry: What's sort of interesting, is that, it would permit two very different client side worlds to share some space
[9:59] Zha Ewry: It has huge scale challanges
http://slurl.com/secondlife/SLVEC/190/144/22 (Starts now)
http://slurl.com/secondlife/Ambleside/181/99/30 (Starts now)
[10:00] Meadhbh Oh: great. they should propose a charter and a working group for it
[10:00] Morgaine Dinova: I pointed out to his that OLIVE doesn't scale if all nodes have to simulate all worlds, and he agreed. But his only escape from non-scalability is by not accepting world diversity, so that all nodes can do the same simulation. He's grudgingly accepting that. But it makes OLIVE a small subset of the VW space, which deals with simulation diversity.
[10:00] Zha Ewry: So... I'm going to try and write up the "co-simulation" idea, and some problems in it
[10:01] Zha Ewry: to get on the list, so people will have a clue about that cornero f things
[10:01] Zha Ewry: and.. what I find interesting, is that I could express the content I'd flow across that
[10:01] Meadhbh Oh: but what does it have to do with the current work of the BoF?
[10:01] Zha Ewry: connection, in the same streaming formet I'd abstractly use to talk sim-> any observer
[10:01] Zha Ewry: or any state injector -> sim
[10:02] Meadhbh Oh: the existing charter is limited in scope and the IETF takes a dim view on research projects
[10:02] Morgaine Dinova: Considering the VW problem space has everything to do with the BoF.
[10:02] Meadhbh Oh: no, it doesn't
[10:02] Morgaine Dinova: Zha?
[10:02] Zha Ewry: I'm goign to split the difference hee
[10:02] Meadhbh Oh: the BoF has to do with interoperability between existing VWs
[10:02] Zha Ewry: *here
[10:02] Meadhbh Oh: not with creating new interaction models
[10:02] Zha Ewry: The IETFS does take a very dim view of research
[10:02] Zha Ewry: Or more tot hep oint
[10:03] Zha Ewry: it seperates it out in seperate spots, where it belongs
[10:03] Zha Ewry: The general goal of ITEF WGs is to usher IETF draft documents through the consensus building, editing, and bakeoff process
[10:03] Zha Ewry: THAT said
[10:04] Meadhbh Oh: right. so the stuff we're talking about WRT to interop with forterra should probably happen outside the IETF, then when there's something resembling a wire protocol, we go back
[10:04] Zha Ewry: The goal of this BOF, and (most likely the severalm onths to the next BOF) is to get a good charter which is supported by enough people in the WG and the IETF that
[10:04] Zha Ewry: the IESG and IAB and IETF wise heads will say "Yeah, verrily, go forth and see if you can get a draft RFC or three into the pipelien"
http://slurl.com/secondlife/SLVEC/190/144/22 (Started 5 minutes ago)
http://slurl.com/secondlife/Ambleside/181/99/30 (Started 5 minutes ago)
[10:05] Meadhbh Oh: mmm... the stated objective for the BoF was to in part, do a "call for participation"
[10:05] Zha Ewry: Call for particpatoin, and bash charter
[10:05] Meadhbh Oh: i don't see that wedging LESS into MMOX is going to be viable
[10:05] Morgaine Dinova: "Produce one or more Internet Drafts, describing the problem
and/or related work. It cannot be emphasized enough that for
the BOF, drafts relating to understanding the problem space are
much more valuable than drafts proposing specific solutions.
"
[10:05] Zha Ewry: What we've gotten is a, to be honest, very limited responde
[10:05] Morgaine Dinova: [2]
[10:06] Meadhbh Oh: morgaine, we're talking about communicatioins between myself, zha and Lisa
[10:06] Zha Ewry: Morgaine's comment is entirely in scope, tho
[10:06] Meadhbh Oh: and while i love LESS as much as the next person
[10:06] Meadhbh Oh: there are NO systems that implement it
[10:06] Meadhbh Oh: at all
[10:06] Zha Ewry: Agreed
[10:06] Zha Ewry: nor is there a community which will bake it off
[10:06] Meadhbh Oh: which makes me think it's a spoiling move
[10:07] Meadhbh Oh: but
[10:07] Meadhbh Oh: if forterra wants to play like an adult
[10:07] Meadhbh Oh: we're open to the idea of trying it again
[10:07] Zha Ewry: notes that would be a statement of Linden Lab, btw
[10:07] Meadhbh Oh: rigth. i'm wearing my linden hat
[10:07] Zha Ewry: nods
[10:07] Zha Ewry: Which is quite expected)
[10:07] Zha Ewry: So...
[10:08] LathaSerevi: I: understand the goal of a BOF meeting to be to demonstrate a consensus that: there is a problem that needs solving, the scope of the problem is well defined and understood,
[10:08] LathaSerevi: there: are people to do the work, and the specific proposed deliverables are the right set.
[10:08] Morgaine Dinova: I don't understand the comment about Forterra "playing like an adult". Jon has submitted an Internet draft. That's precisely what the IETF requires at this point in the process.
[10:08] LathaSerevi: Seems: like Jon is doing that, but Meadhbh doesn't want to.
[10:08] Zha Ewry: I want to touch on two other quick issues,a dn then dive into this discussion a bit deeper
[10:08] Zha Ewry: (Forterra, was exceedingly non constructive at the last pass at this Morgaine)
[10:09] Zha Ewry: Jon's been much more engaged, and to my eye, less difficult, this time around, but.. there are some rather unhappy people all around from the VWIF joyride a year and change ago
[10:09] Meadhbh Oh: Latha, um... i did actually submit a draft (or two)
[10:09] Zha Ewry: My count, is we have 4 or 5 formal documents
[10:09] LathaSerevi: We: have a dramatic non-consensus that the current set of proposed deliverables are the right set. If Meadhbh keeps insisting that her drafts are the only correct set, withou
[10:09] LathaSerevi: t: justifying that, the BOF will go down in flames.
[10:10] Meadhbh Oh: did John H's doc go through?
[10:10] Zha Ewry: I haven't seen it
[10:10] sensory Hax: saluta tutti scusate una cosa urgente spero in un post
[10:10] [[User:xsensory Hax[it>en]|xsensory Hax[it>en]]]: / me salutes all scusate something urgent I hope for a post
[10:10] Zha Ewry: There was a critique, was that submitted as an IETF input?
[10:10] Morgaine Dinova: I'm not party to Forterra's previous previous actions. On the current one, Jon is acting properly, and that's all that matters to MMOX. Don't bear a grudge.
[10:10] Meadhbh Oh: FWIW... our intention is to submit LLSD, OGP/Base, OGP/Auth and OGP/Teleport
[10:11] Zha Ewry: So.. by IETF rules, anything not there today, is not up for deep discussion at the BOF. ANything in by March 2nd, is assumed to have been read by attendees
[10:11] Meadhbh Oh: Zha... which critique?
[10:11] Meadhbh Oh: i think we have til COB tomorrow
[10:11] Zha Ewry: John's detailed tear down of LLIDL and LLSD
[10:12] Meadhbh Oh: cause of problems with the IETF draft submission tool
[10:12] Zha Ewry: Which, had some definite reality behind it.. and is clearly inscope
[10:12] Zha Ewry: Ahh.
[10:12] Zha Ewry: Joy.
[10:12] Meadhbh Oh: sure
[10:12] Zha Ewry: can't say she's surprised, the IETF tools were uber cranky last night
[10:12] Dahlia Trimble: lol
[10:12] Meadhbh Oh: mmm.. i'll ping John and make sure i get a copy of it so it can at least inform the discussion at teh f2f
[10:12] Zha Ewry: So.. two quick things to highlight, and then into the gutys of this
[10:13] Zha Ewry: The Uber Client..
[10:13] Meadhbh Oh: if it for any reason can't be considered "official input"
[10:13] Zha Ewry: Which we've mostly covered in passing..
[10:13] Zha Ewry: Which would be the one client, which could connect to all the VWs out there
[10:13] Zha Ewry: Which, tends to be postulated by people who try to think of this as web pages, needing simple plugins
[10:14] Morgaine Dinova: All "IETF Contributions" are considered "official input". I can paste the paragraphs from the appropriate IETF docs if you want.,
[10:14] Zha Ewry: I think it's pretty unlikely in this decade we'll see anything close to one
[10:14] Meadhbh Oh: @morgaine. that's actually not entirely true
[10:14] Zha Ewry: The issue, tho, is that, at a face to face any material you want to formally consider needs to be on the IETF draft repository
[10:14] Zha Ewry: 3 weeks prior to the start of the meeting
[10:14] Zha Ewry: To give people time to read and digest
[10:15] Zha Ewry: Which is to say, anything we want to talk about, we can
[10:15] Zha Ewry: We can expect people will have actualyl read those documents and be prepared to discuss them
[10:15] Meadhbh Oh: when there's time i'll tell you the story of how RSA (my former employer) abused the internet draft system
[10:15] Rex Cronon: such a client might be possible using addons:)
[10:15] Dahlia Trimble: aren't there commercial interests developing "uber clients" now? I thought I remember seeing references to a few in some recent group chat
[10:15] Zha Ewry: Well, for small overlapping sets of spaces, maybe
[10:15] Meadhbh Oh: are there?
[10:15] Zha Ewry: In the large, it would be insanely difficult
[10:16] Dahlia Trimble: there was one that could let you play any online game in a web browser
[10:16] Zha Ewry: For example, to do croquet based stuff you need the full t-object framework
[10:16] Meadhbh Oh: when we say "uber-client" do we mean one that connects to EVERYTHING
[10:16] Meadhbh Oh: or one that connects to MANY thigns?
[10:16] Zha Ewry: to do Olive you would need to be able to do the whole "lock step physical simulation"
[10:16] Zha Ewry: to do SL, you need to do the SL dead-reckoner, and things like the avatar client side stuff
[10:17] Dahlia Trimble: many things would seem the likely approach
[10:17] Meadhbh Oh: ohoh dalia... you may be thinking of the people who distribute raster lines to web clients
[10:17] Meadhbh Oh: which is really cool stuff
[10:17] Rex Cronon: i don't think is possible to have one client that connects to all VWs at the same time
[10:17] Morgaine Dinova: Difficult, but not impossible, Zha. After all, web browsers connect to "everything". That's not possible to do with Jon's OLIVE model, because he does simulation in all clients. But it's very possible in the SL/Opensim model, since simulation is left to the regions, and the clients only get telemetry back.
[10:17] Zha Ewry: Torque uses a totally different client side avaatr scheme, as does, last I look IMVUU
[10:17] Meadhbh Oh: but i would argue is different 'cause the web browser isn't a protocol endpoint for the games being played
[10:18] Zha Ewry: Its pretty daunting, at best
[10:18] Zha Ewry: So.. the reason I bring this up is
[10:18] Dahlia Trimble: I think the one I saw was running the game as a separate process and somehow displaying the opengl in a browser
[10:18] Meadhbh Oh: morgaine... web browsers connect to everything because HTTP is $DIET{IES|Y}'s own protocol
[10:18] Dahlia Trimble: but i didnt look too closely
[10:18] Zha Ewry: and they only show you what you have plugins for
[10:18] Zha Ewry: which mostly works, because, worst case, you can pipe the pile of bits to a process and let it handle it
[10:19] Morgaine Dinova: All that generic clients will need to support if a diversity of object standards, and in principle the set can be quite dynamic, with worlds making the extensions available on demand. Hard, but not impossible.
[10:19] Zha Ewry: (Note the more you do AJAX like stuff, the less it works everywhere)
[10:19] Zha Ewry: Anyway...
[10:19] Zha Ewry: Tha's what I mean by "uber clients"
[10:19] Zha Ewry: and.. the reason I bring this up is that
[10:19] Meadhbh Oh: right... the plugin model for browsers has the browser give the plugin a bit of window real-estate and says... knock yersefl out
[10:19] Zha Ewry: for quite a while.. pragmatic interop is going to be in much smaller sets spaces
[10:20] Meadhbh Oh: definitely
[10:20] Dahlia Trimble: it may have been a glorified glintercept or something
[10:20] Zha Ewry: And.. what Linden has proposed here.. which is one part of interop, is to do interop in a set of broadly similar spaces which share an overall client approach
[10:20] Meadhbh Oh: right. my point is the browser doesn't need to understand the model for the plugin's data
[10:20] Zha Ewry: and, grow out from there
[10:20] Meadhbh Oh: the plugin understand that
[10:20] Meadhbh Oh: right... what i call "model 5"
[10:20] Zha Ewry: (nor do plugins cooperate)
[10:21] Rex Cronon: plugins could cooperate
[10:21] Zha Ewry: The problem with the "plugin model" for VWs, is things like "Do I need a plugin to handle inventory in 5 different worlds" as well as ones to render it
[10:21] Morgaine Dinova: Also, note that Limited Capability Clients can be multiple world clients easily, since they need compatibility across only a small part of the feature set. Eg. there's no reason at all why a screenless Braille client shouldn't connect to a dozen worlds, for text-to-Braille only.
[10:21] Zha Ewry: nods
[10:21] Zha Ewry: Which Jon didn't like at all Morgaine ;-)
[10:22] Meadhbh Oh: so yeah. in that sense jon is correct. we assume a few things about the virtual world... that machines that do simulation may not be the machines that have authority over the data used as inputs into the simulation
[10:22] Meadhbh Oh: and for the near term, yes, we are going to assume that all virtual worlds will want to consider the option of teleporting
[10:22] Morgaine Dinova: Not surprised he didn't like it. I'm trying to tread a middle line. :-)
[10:22] Zha Ewry: Actually, Meadhbh, I'd say it differently
[10:23] Zha Ewry: "For now we're going to assume there is value in doing a set of worlds which share a lot of common modeling, such as teleport"
[10:23] SharedRealm Engineer: is one of the goals of MMOX to define the client<->sim protocol?
[10:23] Zha Ewry: Not at all clear, to me
[10:23] Zha Ewry: Not first up for sure
[10:24] Meadhbh Oh: @Shared. yes. client <-> server and server <-> server are considered appropriate for discussion
[10:24] Morgaine Dinova: SharedRealm: the issue there is the word "sim", in part. If you say "service endpoint", you'll get much more agreement. :-)
[10:24] SharedRealm Engineer: If you don't have an uber client then you need one protocol for client<->sim comms.
[10:24] Meadhbh Oh: right. in our specs we don't say "sim"
[10:24] Zha Ewry: I tend to go even more abstract
[10:24] Meadhbh Oh: or... at least we shouldn't
[10:24] Zha Ewry: "state injector"
[10:24] Zha Ewry: "User represenation"
[10:24] Zha Ewry: "Observer viewpoint"
[10:25] Meadhbh Oh: ugh. "state injector" is, i believe, a forterra trademarked term
[10:25] Zha Ewry: Because that lets you tease away all the current implementtaion couplling and look at the services you need
[10:25] SharedRealm Engineer: then you only need one client to connect to all the interop. VWs.
[10:25] Teravus Ousley: Additionally, current web clients have some standard controls and interactions... hyperlinks, forms.. submit buttons.. these are all researched as some of the most effective ways to get input from users. And, users have been conditioned to know what they mean.. in the 3D space, we don't have those sets of standards.
[10:25] Zha Ewry: You have to be kidding me Meadbh
[10:25] Zha Ewry: "State" has been "injected" into simulations for decades
[10:25] Meadhbh Oh: i
[10:26] Meadhbh Oh: i'm just saying... from the VWIF days... it seemed like a lot of people were saying a lot of very, very stupid things
[10:26] Meadhbh Oh: but yeah... teravus... +1
[10:26] Teravus Ousley: .. and therefore, we can't say.. in the 3D wire protocol.. we need a 'touch' action on this object.. and a 'web browser' represented on this face.. and..
[10:26] Meadhbh Oh: i might argue that web forms weren't really researched
[10:26] Meadhbh Oh: but
[10:26] Meadhbh Oh: they work
[10:26] Dahlia Trimble: lol
[10:26] Dahlia Trimble: lol
[10:27] Meadhbh Oh: and people know what they mean... what do the UCE peeps call em? affordances?
[10:27] Zha Ewry: right, and in fact, the reason I tend to seperate state injection from the rest
[10:27] Teravus Ousley: We're in the 'flash interface' state of the 3D space. Flash interfaces are often looked down upon in the 2d Design space because they're non-standard.
[10:27] Zha Ewry: is that.. they are lessunderstood
[10:27] Zha Ewry: and, yes
[10:27] Zha Ewry: affordances
[10:27] Meadhbh Oh: ahah... k. now i get your context
[10:27] Zha Ewry: Which is a broader term
[10:27] Zha Ewry: since we could say "the ability to drag a PDF onto a prim"
[10:27] Zha Ewry: is an affordance in UCE terms
[10:28] Meadhbh Oh: definitely doesn't want to get in another argument about the differences between affordances and artificies
[10:28] Morgaine Dinova: Tera: we're treading new ground in the interop space, so hard to judge. But people like common clients, because they're used to web browsers. So my guess is that the open source community will head in that direction, and in due course, that will cause massive convergence.
[10:28] Zha Ewry: works *very* closely with our User Centered Design team
[10:28] Zha Ewry: So...
[10:28] Meadhbh Oh: i would argue that people like tools that let them do their jobs
[10:28] Meadhbh Oh: though
[10:28] Zha Ewry: Two other things which I wanted to touch on from the mailling list
[10:28] Meadhbh Oh: if you could do your job from within a single tool
[10:28] Meadhbh Oh: great!
[10:28] Zha Ewry: before we continued to figure out how to sort out the "What the heck are we going to work on" tangle
[10:29] Teravus Ousley: It would look like 3D studio Max.. as far as interface.. haha
[10:29] Zha Ewry: says "Teravus, I'm pretty sure I can drive to where you are physically sitting and remove a few more organs"
[10:29] Rex Cronon: if somebody makes a plugin for it , ucould:)
[10:29] Zha Ewry: is not a fan of 3dMax UI
[10:29] Teravus Ousley: :D
[10:30] Zha Ewry: sees double and shakes her head hard
[10:30] Teravus Ousley: that was why I brought it up. I'm not particularly a fan either. It's just.. it encapsulates an OMG amount of functionality.
[10:30] Zha Ewry: Meadabh, if the two of you start necking, I'm going to get weired out
[10:31] Meadhbh Oh: lol
[10:31] Zha Ewry: So..
[10:31] Rex Cronon: i find it hard to belive that 3dmax ui is worse than what blender has:)
[10:31] Infinity Linden: nah. we do that in private
[10:31] Zha Ewry: Two other things I noticed recently
[10:31] biskit covers: his eyes
[10:31] Zha Ewry: There are a "lot" of "Persistent named places" assumptions in our thinking
[10:32] Zha Ewry: I tossed some of the MMOX and OGP stuff towards one of our lotus people working on ad-hoc 3d meeting spaces
[10:32] Zha Ewry: andf they were like "Wow, you'r all over places which are pesistent"
[10:32] Zha Ewry: which, yeah, we are
[10:32] Zha Ewry: Doesn't change much of the technology
[10:32] Zha Ewry: But.. its seomthing we may want to be aware of
[10:32] Meadhbh Oh: yeah... but the persistence is based on an assumption by the server and client
[10:33] Meadhbh Oh: i don't think it's a requirement of the wire protocol
[10:33] Zha Ewry: For example, both the croquet and OLIVE design points, assume far smaller amounts of long term persistent world
[10:33] Infinity Linden: likes the concept of ad hoc spaces
[10:33] Teravus Ousley: well, also by definition of 'Virtual world'.. is that it's persistant.. however.. not necessarily for MMOeXperience
[10:33] Zha Ewry: lurves it, to a point
[10:33] Dahlia Trimble: bye all :)
[10:33] Infinity Linden: maybe it's something we could talk to forterra about?
[10:33] Rex Cronon: tc dahlia
[10:34] Zha Ewry: I just mention this, because, its the lead in to my last topic, before we deep dive
[10:34] Infinity Linden: after the name calling... it would be nice to get something / anything that we could work on together
[10:35] Zha Ewry: Beyond persistence, the whole OGP spec set, not surprisingly for its history, is sort of riddled, softly, with assumptions about how the world is composed. They are mostly valid, they tend to be expressed in very specific, and specliazed names,a nd can obscure what we're really talking about
[10:35] Zha Ewry: Jon, was taking way to much about "rez_avatar/request" as being only for injecting an avatar
[10:35] Meadhbh Oh: yup. but. OGP was designed with web access in mind
[10:35] Zha Ewry: when.. in fact, you could get the set of caps from a sim, and then use them differently
[10:36] Meadhbh Oh: or at least we added a few bits that only make sense from a web access perspective
[10:36] Teravus Ousley: Are you comparing OGP language to that of Jon's comment about the 'language of forterra' last week?
[10:36] Zha Ewry: Nope
[10:36] Morgaine Dinova: Another built-in assumption seems to be the point of authority for objects, which is very server-centric in our proposed models so far. Not sure how to tackle that, but certainly we don't want to build too many architectural assumptions in.
[10:37] Zha Ewry: Just observing that one thing we (and this is thw AWGroupies, OpenSimers, and Lindens, and all of us who have been soaking in this brine for several years)
[10:37] Morgaine Dinova: We had a good discussion with Jesrad about that.
[10:37] Meadhbh Oh: yes. OGP is a server centric protocol
[10:37] Meadhbh Oh: but there is nothing in the protocol that mandates you talk to "a simulator"
[10:37] Zha Ewry: tend to do, id focus on a very speciric set of terms, because they are useful, but, in fact, they are one factoring of the services
[10:37] Morgaine Dinova: Only if you make it one. OGP is based on endpoints, which are more flexible than SL.
[10:38] Zha Ewry: So, you may recall, 12 months ago, when Zero puhed on Agent Domain as having seperate endpoints for some services
[10:38] Teravus Ousley: Ah, to summarize, "Think outside the box"?
[10:38] Meadhbh Oh: fwiw... there's a reason OGP is server centric. both LL and OpenSim are server-centric
[10:38] Zha Ewry: and.. a lto of peopel freaked
[10:38] Rex Cronon: u could assume that eventually all VW will have their users store personal data locally:)
[10:38] Zha Ewry: My point, is that, with no loss of function, and a big gain in comprehensino, we could make that more clear
[10:38] Meadhbh Oh: right. but do you really want everyone in a virtual area to be hammering on your DSL uplink like that?
[10:38] Morgaine Dinova: And it's that SL+Opensim-only thing that MMOX is there to remove. That's the whole idea of MMOX, to go beyond the first model.
[10:38] Zha Ewry: "This is the set of steps used to get access to a virtual space simulator"
[10:39] Zha Ewry: its the same as "Thsi is how you do a rez_avatar/request"
[10:39] Meadhbh Oh: a LOT of networks (especially ones that are offered to home users) are asymmetric in terms of direction
[10:39] Rex Cronon: that data could be uploaded to the sim u login
[10:39] Rex Cronon: similar to the new way textures r supposed to work:)
[10:39] Meadhbh Oh: which is one of the reasons LL's solutions are server centric
[10:40] Meadhbh Oh: (cause remember... Infinity and I come from the Smalltalk / Squeak / Croquet world)
[10:40] Teravus Ousley: No, we don't really want to hammer a DSL uplink, however.. we will right for your right to do that if you want to ! .. by making the definition generic enough that you can use whatever quality network you desire.
[10:40] Teravus Ousley: .. "fight for your right"
[10:40] Meadhbh Oh: (there had to have been a number of good reasons to move us away from the distributed t-object model)
[10:40] Zha Ewry: Note, tho, if you get the model clean, the easiest way to do this, is to have your assets stored in a local asset server, and have an interop level relationship between *that* (not the client) and the sim
[10:41] Zha Ewry: If your assets are all cleanly referred to with URIs
[10:41] Zha Ewry: and you are willing to host an asset server?
[10:41] Zha Ewry: Go to town
[10:41] Meadhbh Oh: but... the idea that the server you talk to may not be the authoratative repository for personal info is definitely interesting
[10:41] Meadhbh Oh: but
[10:41] Rex Cronon: a user could possible rent a webspace to store his/hers own data, so no need to hammer the poor dsl:)
[10:41] Zha Ewry: (and buy a fat pipe from your ISP)
[10:41] Morgaine Dinova: Tera: indeed. There's no reason why a noddy PC world on DSL can't support huge crowds .... if there's a suitable commercial big-iron caching proxy in front, as a paid service. ;-)
[10:41] Meadhbh Oh: i think including that in the client <-> server protocol is not a bag of win
[10:41] Teravus Ousley: Amazon S3! :D
[10:41] Morgaine Dinova: Yep!
[10:42] Zha Ewry: nods at Rex (And that's again, exactly, why the model is "URI->Service", not "in the client")
[10:42] Meadhbh Oh: @Morgaine... what color is the grass in your world?
[10:42] Morgaine Dinova: The colour of grass in my world is off topic
[10:42] Teravus Ousley: my grass is purple.. .. kind of hides me... considering my color..
[10:42] Meadhbh Oh: yeah. mine is actually red.
[10:42] Zha Ewry: hands Morgaine more neon pink paisely virtual paint.
[10:42] Meadhbh Oh: but you're right
[10:42] Meadhbh Oh: it's off topic
[10:42] Morgaine Dinova: Tnx Zha
[10:43] Zha Ewry: So..
[10:43] Rex Cronon: so what is the problem?
[10:43] Zha Ewry: Here we go
[10:43] Zha Ewry: Given all of this
[10:43] Zha Ewry: we have two or three things we'd like to manage
[10:43] Zha Ewry: I for one, (and my corporate masters)
[10:43] Meadhbh Oh: i think my point was... "we don't need to add proxy and cache managemetn to the viewer <-> server protocol"
[10:43] Rex Cronon: the user just needs to provide an address of where the data is located:)
[10:43] Zha Ewry: want to see OGP like stuff advanced, in as public a square as possible
[10:44] Zha Ewry: Growing out the ecosystem which is developing around second life, is, pretty much a good thing
[10:44] Meadhbh Oh: in fact... i would argue that giving viewer "a" the ability to say "okay... take object 1 from viewer "b"s machine and put it on server Z" is a mistake
[10:44] Rex Cronon: u could say that the user send the data location as part of the connection protocol:)
[10:44] Zha Ewry: (and growing it out in a way that has standards, and a decent decomposition, and such is a good thing)
http://slurl.com/secondlife/Denby/213/45/34 (Starts in 15 minutes)
[10:45] Meadhbh Oh: likes the idea of having all viewers participating in the consentual hallucination talking to a limited number of high-bandwidth servers
[10:45] Zha Ewry: Well, if server Z offers up a cap to store Item X, the user can always push it here
[10:45] Zha Ewry: (one of the nice things is that becomes a policy, not a protocol question)
[10:45] Zha Ewry: OHHHH
[10:45] Zha Ewry: and that was the last thing I wanted to talk about before the deep dive
[10:45] Meadhbh Oh: and then having those servers have different agreements with each of the clients in terms of how they manage pulling authoratative data from them
[10:46] Morgaine Dinova: There's plenty of precedent for service redirection. Redirection for resources is perfectly mainstream for REST. :-)
[10:46] Rex Cronon: isn't a policy part of the protocol?
[10:46] Zha Ewry: I made an I think slightly controversial, but very important comment, in the scope/problem writeup I posted
[10:46] Zha Ewry: NEWP
[10:46] Zha Ewry: The affordances to enable policy choices are
[10:46] Meadhbh Oh: morgaine.. no. rex... no.
[10:46] Zha Ewry: the policies, are a totally seperate concern
[10:46] Meadhbh Oh: but let's talk about it more later
[10:47] Meadhbh Oh: i think zha's trying to move us to a different agenda item
[10:47] Zha Ewry: So, we need to make sure we put in all the annoying things like "Here is the user's proof of who they are"
[10:47] Zha Ewry: into the protocol
[10:47] Rex Cronon: excuse me, but isn't the protocol describing how to connet?
[10:47] Meadhbh Oh: no
[10:47] Meadhbh Oh: it describes mechanisms to connect
[10:47] Zha Ewry: so, that various deployers/users of the protocol can then have policies
[10:47] Morgaine Dinova: We're defining only mechanisms, not policy. If a VW provider has a policy of not supporting asset redirection, that's its business. But it's certainly par for the course for the mechanism to allow it.
[10:47] Meadhbh Oh: +1 morgaine
[10:48] Zha Ewry: So, the mechanism says "Here is how you push an asset to a service"
[10:48] Zha Ewry: and the policy says
[10:48] Infinity Linden: and for the record, we don't give client A's IP address to client B.
[10:48] Zha Ewry: "You need to have a one time token in your request granted from me, from my back channel to use it"
[10:48] Infinity Linden: it's considered a no-no
[10:48] Meadhbh Oh: but for a world that's totally behind a corporate firewall
[10:48] Meadhbh Oh: it's probably not a big deal
[10:48] Meadhbh Oh: so
[10:49] Meadhbh Oh: it's okay to go into a protocol, but not okay to mandate it's use
[10:49] Zha Ewry: So, in general, we totally divorce policy from the mechanism, with *ONE* tiny exception
[10:49] Zha Ewry: Which is making sure we leave the spots in the protocol to put things policy users need
[10:49] Morgaine Dinova: Yep
[10:49] Zha Ewry: Note, this is very, very, webish
http://slurl.com/secondlife/Denby/213/45/34 (Starts in 10 minutes)
[10:50] Zha Ewry: also note, it totally borks 90% of static content schemes
[10:50] Meadhbh Oh: bites tongue... she was about to mention that plugins are also considered webbish
[10:50] Zha Ewry: Its wickedly hard to get the "my totally private one time use token" formatte dinto a public spec
[10:51] Zha Ewry: but, wickedly easy, to say "Oh, in the optinoal elements map of a dynamic map scheme, slip in "Zha's one time token, map of ints"
[10:51] Meadhbh Oh: mmm.. in what sense?
[10:51] Meadhbh Oh: ohoh
[10:51] Meadhbh Oh: yup. CORBASec 2 was even not a manditory extension
[10:52] Zha Ewry: Classic, old school, XML isn not really friendly to dynamic extensions of elements in yoru scheme tree
[10:52] Meadhbh Oh: though... we consider the use of caps as being an "essential" part of our operating environment
[10:52] Zha Ewry: brb
[10:52] Zha Ewry: (about 2-3 mins)
[10:52] Meadhbh Oh: cries thinking about the ways XML has been abused
[10:52] Infinity Linden: hands meadhbh a tissue
[10:53] Meadhbh Oh: so... how bout that local sports team?
[10:54] Meadhbh Oh: as i suspected.... everyone else here is one of Zha's alts!
[10:54] Infinity Linden: shhh! not so loud!
[10:54] Rex Cronon: i am not
[10:54] Areal Loonie: ha
[10:54] Teravus Ousley: shh, don't tell !
http://slurl.com/secondlife/Denby/213/45/34 (Starts in 5 minutes)
[10:55] Zha Ewry: skims back chat as she returns
[10:55] Rex Cronon: testing...
[10:56] Zha Ewry: checks for alts and doesn't find any of hers
[10:56] Rex Cronon: and i thought i was disconnected:)
[10:56] Meadhbh Oh: whistles to herself
[10:56] Infinity Linden: as does i
[10:56] Morgaine Dinova: Better find the next topic Zha:-)
[10:56] Zha Ewry: So...
[10:56] Zha Ewry: Couple of things which I think would help
[10:57] Zha Ewry: one.. we need to both set the scope broadly enough, in the preface of the charter, and then justify why we're working on one part of the space. I think that's actually pretty easy.
[10:57] Zha Ewry: "Look, this is where there are multiple public overlapping implementations, that's what we can spec"
[10:57] Zha Ewry: But..
[10:58] Zha Ewry: We should not be afriad, to for example, encourage Jon LESS is more approach, if he can
[10:58] Meadhbh Oh: agree'd
[10:58] Meadhbh Oh: but
[10:58] Zha Ewry: for example show an OpenSim/LESS/OLIVE gateway
[10:58] Meadhbh Oh: LESS is not implemented
[10:58] Meadhbh Oh: anywhere
[10:58] Zha Ewry: then... and only then.. do we get to write specs about it
[10:58] Meadhbh Oh: and if i'm wrong then i'll DEFINITELY take a different stance
[10:58] Zha Ewry: nods at Meadhbh
[10:59] Zha Ewry: (I have to put that on a gesture, the hbh is just awful to type)
[10:59] Meadhbh Oh: 'cause my superiors at the IETF will let us move forward with things that define the behavior of existing or soon to be implemented systems.
[10:59] Morgaine Dinova: It doesn't have to be implemented, at this stage of the IETF process. If IETF standards only discussed what was implemented before, there would be nothing new. :-)
[10:59] Meadhbh Oh: but
[10:59] Meadhbh Oh: if jw's gonna implement a LESS gateway
http://slurl.com/secondlife/Denby/213/45/34 (Starts now)
[11:00] Meadhbh Oh: that is actually very, very coll
[11:00] Meadhbh Oh: cool
[11:00] Zha Ewry: well, the IETF is funky that way, it doesn't do a lot of invention
[11:00] Meadhbh Oh: and we want to encourage it
[11:00] Zha Ewry: It captures specs as inovaction congeels into usable code
[11:00] Morgaine Dinova: Indeed. And it's only AFTER that that the refernece implementations get to work.
[11:01] Zha Ewry: Actually, no
[11:01] Zha Ewry: The most common pattern at the IETF is
[11:01] Meadhbh Oh: yeah. the ietf doesn't do reference implementations
[11:01] Zha Ewry: that peopel come with one or more workiing systems
[11:01] Zha Ewry: and, then they bash the desciprtion of them into a spec
[11:01] Zha Ewry: and refine the systems to match
[11:01] Zha Ewry: and bakeoff those
[11:01] Zha Ewry: and use that to validate the spec
[11:01] Morgaine Dinova: The IETF doesn't *DO* reference implementation,s but it requires them to be done by others.
[11:01] Zha Ewry: Not exactly, morgaine
[11:02] Zha Ewry: its a subtle distinction
[11:02] Meadhbh Oh: has bad, but somehow good memories of the IPSec bakeoffs
[11:02] Zha Ewry: The IETF makes the documents normative, not any one implementation
[11:02] Meadhbh Oh: no. there are NO reference implementations
[11:02] Zha Ewry: I used to think that this was a weakness
[11:02] Zha Ewry: BUT...
[11:02] Zha Ewry: Having had to live the nightmare of
[11:02] Zha Ewry: "But, that badly formed packet is accepted by the Foogleblaz reference implementation"
[11:02] Zha Ewry: when it cleary didn't match the spec
[11:03] Infinity Linden: LOL.
[11:03] Zha Ewry: I'm less sympathetic to the reference implementation approach
[11:03] Zha Ewry: mutters
[11:03] Infinity Linden: laughs at the thinly veiled reference to MICO
[11:03] Zha Ewry: The spec clerarly says "Netwrok endian"
[11:03] Rex Cronon: time 4 me 2 go 2 andrews office hour
[11:03] Rex Cronon: bye everybody
[11:03] Teravus Ousley: [sic]
[11:03] Zha Ewry: By rex
[11:03] Infinity Linden: cheers!
[11:03] Rex Cronon: tc:_
[11:03] Zha Ewry: So.....
[11:03] Rex Cronon: :)
[11:04] Zha Ewry: I'm suggesting we tweak our charter a tiny bit, to both make the scope clearer (and a little broader) and then explicitly say "At this point in time, we can work on OGP and such as early MMOX workproducts"
http://slurl.com/secondlife/Denby/213/45/34 (Started 5 minutes ago)
[11:05] Zha Ewry: "We'll gladly work on IETF drafts in orher parts of the space, as they become mature"
[11:05] Teravus Ousley: so, what are we going to discuss at the Zero meeting?
[11:05] Teravus Ousley: .. what Drafts?
[11:05] Zha Ewry: And, we'll try and get as many Sl centric wording out of our specs as we can, and keep the abstractions broader
[11:05] Teravus Ousley: .. what items are we going to bring up?
[11:05] Infinity Linden: and for what it's worth... there _are_ zero office hours today, but apparenly no one suggested agenda items. so we might talk about the process of getting items on the agenda
[11:06] Zha Ewry: zero agenda at Zero
[11:06] Infinity Linden: @terevus
[11:06] Infinity Linden: LLSD, LESS, OGP/Base have been submitted
[11:06] Infinity Linden: i haven't looked to see if JHurliman has submitted his...
[11:06] Zha Ewry: A problem./scope document to help ITEF peopel find the way is in there
[11:06] Zha Ewry: if in fact, I have 48 hours
[11:06] Zha Ewry: I *may* try to add one more
[11:07] Infinity Linden: and i'm rushing to get OGP/Auth and OGP/Teleport submitted so they can be discussed @ the f2f meeting
[11:07] Zha Ewry: would love an informative summary of a bunch of topics
[11:07] Zha Ewry: including my "in my head, and somewhat discussion on MMOX chat" full decomposition of
[11:07] Zha Ewry: the Sim/Client connection into its constituent parts
[11:08] Morgaine Dinova: Zha: It's worth noting that there is no document currently justifying use of the OGP model. In AWG we know the justifications because we've been working on it for 16 months, but that's not enough for MMOX. I suggest that either in your revised charter, or in a separate document, the technical justification for choice of OGP for part of the MMOX space be made, just to put the MMOX side of the work on it on solid technical footing.
[11:08] Meadhbh Oh: yup. at the very least a discussion that says... "yeah... content negotiation didn't make it into OGP/Base"
[11:08] Zha Ewry: I agree
[11:08] Meadhbh Oh: and here's how we think ig should be done
[11:08] Zha Ewry: its a weakness in the whole story
[11:08] Zha Ewry: For example
[11:09] Meadhbh Oh: @morgaine. yes. there is no such document
[11:09] Zha Ewry: Zero and Infinity have very good reasons why LLSD and not base XML.
[11:09] Meadhbh Oh: though Lisa has addressed this on the list
[11:09] Zha Ewry: And.. to a large extent, they are derived from practical expeirnce
[11:09] Meadhbh Oh: if you want to charter your own working group, you're more than welcome to do so
[11:09] Zha Ewry: But.. they are not documented
[11:09] Zha Ewry: Yah, but, we're not just talking to ourselves
[11:09] Meadhbh Oh: and it may happen that jw and the croquet peeps go off in one direction in one WG
[11:10] Meadhbh Oh: and the OGP peeps go off in another
[11:10] Zha Ewry: the IETF on the whole, may want to understand those chocies
[11:10] Meadhbh Oh: it's not a failure
[11:10] Zha Ewry: Infinity, I'm not opposed to that, but.. I'm also not suporting it, unless someone shows me why we can't keep the bits of work in one WG.
[11:10] Meadhbh Oh: right
[11:10] Zha Ewry: There is huge overlap between thinsg like mimetypes for content
[11:11] Zha Ewry: and managing access to stuff
[11:11] Morgaine Dinova: Zha, read your PM please.
[11:11] Zha Ewry: and names for things, and so on
[11:11] Zha Ewry: (did)
[11:11] Morgaine Dinova: nods
[11:11] Zha Ewry: So.. teling people "We have to chase you away"
[11:11] Zha Ewry: is not terribly constructive
[11:11] Meadhbh Oh: ack
[11:11] Meadhbh Oh: yes
[11:11] Zha Ewry: We don't want to NOT do the OGP work
[11:11] Zha Ewry: but that's different from only doing the oGP work
[11:12] Meadhbh Oh: but when people come up and say.. "we're going to block your progress 'cause we have private reasons to not want a spec in this space and we're not going to deploy a system that uses an alterntative."
[11:12] Meadhbh Oh: well
[11:12] Zha Ewry: I'd rather keep crticial mass in one place, and push the various people in the other branches of the VW cloud overlapped where we can
[11:12] Meadhbh Oh: i don't think that works either
[11:12] Zha Ewry: That's different
[11:12] Zha Ewry: And. when someone says "No, you can't work on OGP" then, we have that fight
[11:13] Meadhbh Oh: umm...
[11:13] Zha Ewry: But. if someone says "MMOX should be bigger than JUST OGP"
[11:13] Meadhbh Oh: i think jw has already said that
[11:13] Meadhbh Oh: right
[11:13] Zha Ewry: Nah.
[11:13] Meadhbh Oh: which is what i've been saying
[11:13] Zha Ewry: Jon's not saying that
[11:13] Meadhbh Oh: but
[11:13] Zha Ewry: He's saying 'You can't call it all of MMOX"
[11:13] Meadhbh Oh: okay.. as long as he says... "i have no problem with OGP being in the charter"
[11:13] Meadhbh Oh: i'll believe it
[11:13] Meadhbh Oh: cause
[11:13] Zha Ewry: and, as long as we throw him the bone of "This is part of MMOX"
[11:13] Meadhbh Oh: as it stands
[11:14] Zha Ewry: I'm OK.
[11:14] Meadhbh Oh: i have no problem with LESS being on the charter as long as OGP gets to be on the charter
[11:14] Zha Ewry: Amusingly, he reacted most strongly to John Hurliman'srevision of the charter
[11:14] Zha Ewry: Whch took out Linden
[11:14] Meadhbh Oh: LOL
[11:14] Zha Ewry: and put in "vendor neutural"
[11:14] Zha Ewry: which, I think he found threatening
[11:14] Meadhbh Oh: LOL
[11:14] Meadhbh Oh: right
[11:14] Zha Ewry: So..
[11:14] Teravus Ousley: heh
[11:14] Zha Ewry: What I'm inclined to say is:
[11:15] Zha Ewry: MMOX is broad enough to encompass lots of work, and we encourage it
[11:15] Zha Ewry: BUT
[11:15] Teravus Ousley: well, if we want this standard to actually mean something.. it will have to be vendor neutral in the end.. :D
[11:15] Meadhbh Oh: because jon doesn't want to see a spec that allows us to say... "we have a spec that governs interop with SL" <- that's an opinion, btw
[11:15] Meadhbh Oh: not backed up with facts
[11:15] Meadhbh Oh: right
[11:15] Zha Ewry: we have this stuff which is actually in code
[11:15] Teravus Ousley: even if it encorporates technologies from various companies :D
[11:15] Zha Ewry: and seperate code bases
[11:15] Zha Ewry: and we want to get it into an IETF track
[11:15] Zha Ewry: so.. we're doing that
[11:15] Zha Ewry: Go ahead, add LESS to the charter
[11:16] Meadhbh Oh: right. we've already demonstrated that we want to work with multiple organizations and vendors
[11:16] Zha Ewry: LESS is more
[11:16] Zha Ewry: but only if its MORE
[11:16] Zha Ewry: not in place of
[11:16] Zha Ewry: and.. I woud *loev*
[11:16] Meadhbh Oh: but jw's has just published a draft to say... "see... the OGP way is not the only way to do things!"
[11:16] Meadhbh Oh: which is true
[11:16] Meadhbh Oh: but
[11:16] Meadhbh Oh: for IETF work, there has to be an implementation or two
[11:16] Meadhbh Oh: though
[11:17] Zha Ewry: to actually do the unified state transfer model which lets me stream one stream from a sim to a client, and another to a co-simulating sim
[11:17] Meadhbh Oh: if the croquet peeps picked up on LESS
[11:17] Meadhbh Oh: and supported it
[11:17] Morgaine Dinova: Zha: your IETF draft got resounding approval all around for being even-handed and vendor neutral. I think that's a good sign of where MMOX as a group stands. I don't think I've heard anyone want anything but MORE ... even LESS is more. :-) It should simply not be as narrow as SL+Opensim, because that wouldn't get interop very far.
[11:17] Meadhbh Oh: i would see if i couldn't figure a way to get it integrated into our system
[11:17] Meadhbh Oh: so
[11:17] Zha Ewry: nods
[11:17] Meadhbh Oh: you could have Croquet / Forterra interop wiht SL via LESS
[11:17] Zha Ewry: I'm leaning to saying
[11:17] Zha Ewry: "Show me some code, Jon"
[11:18] Meadhbh Oh: it would be A FRICKEN HUGE amount of engineering work, so it's not going to happen in the next couple of years
[11:18] Zha Ewry: and we'll do a unified state transfer model so we can do a bakeoff in 2011
[11:18] Meadhbh Oh: yup. and what i'm saying is i'll help write the code if jon can convince the croquet peeps to adopt LESS
[11:18] Zha Ewry: where we have a Croquet, OLIVE, and a OGP sim
[11:18] Zha Ewry: and we couple them via sim-sim-sim
[11:18] Zha Ewry: and then clients
[11:18] Zha Ewry: and show it all
[11:18] Zha Ewry: And tp from one of thoe to another via OGP
[11:19] Zha Ewry: and then I'll go an buy eveyone who made it work a serious drink
[11:19] Morgaine Dinova: Jon doesn't have to convince the Croquet people to implement LESS. He only has to convince MMOX that it would be a good thing to allow.
[11:19] Morgaine Dinova: Mechanism, not policy
[11:19] Zha Ewry: To get it beyond talking, he had s get code
[11:19] Meadhbh Oh: right... though i imagine croquet and forterra users will always be talking to croquet-LESS or forterra-LESS gateways that would talk directly to OGP sims
[11:19] Zha Ewry: He can submit a draft
[11:19] Meadhbh Oh: (and vice versa with respect to SLish viewers and SLish sims)
[11:19] Zha Ewry: but if he doesn't want it to whiter after 9-12 onths
[11:19] Zha Ewry: he has to actually show code
[11:20] Zha Ewry: Which is frankly, as it should be
[11:20] Zha Ewry: (And, there are some really amusing scale issues)
[11:20] Meadhbh Oh: @morgaine. jon needs to convince the croquet people to use LESS in order to get my personal support
[11:20] Morgaine Dinova: Not relevant to IETF
[11:20] Meadhbh Oh: which would include (but not be limited to) writing a squeak LESS to OGP gateway
[11:20] Zha Ewry: Well, the IETF won't let him advance his draft
[11:20] Zha Ewry: without some working code
[11:20] Meadhbh Oh: @morgaine. yes. very related to IETF
[11:21] Zha Ewry: He can publish as many drafts as he wants, but the ADs, IESG and IAB won't let them advance beyond the early states, without code behind them
[11:21] Zha Ewry: (Just as they won't let us put stuff into the OGP spec which doesn't have code)
[11:21] Zha Ewry: (and, more than one copy of the code, in fact)
[11:22] Meadhbh Oh: from multiple genetically distinct sources
[11:22] Morgaine Dinova: Turn it around to see how poor the suggestion is: what you're suggesting is that for Linden to get OGP into the MMOX spec, then Linden first has to convince OLIVE people to implement OGP.
[11:22] Zha Ewry: however defined
[11:22] Morgaine Dinova: Really quite silly.
[11:22] Zha Ewry: Actually?
[11:22] Meadhbh Oh: NO
[11:22] Meadhbh Oh: this is not true
[11:22] Zha Ewry: We didn't say any specific environment
[11:22] Meadhbh Oh: we have a draft and code
[11:22] Zha Ewry: Just.. two genetically diverse
[11:22] Meadhbh Oh: with demonstrated interoperability
[11:22] Zha Ewry: So, if Jon shows up with OLIVE and qwaq
[11:22] Zha Ewry: or LIVE and IMVU
[11:23] Zha Ewry: or OLIVE and VRML
[11:23] Teravus Ousley: Or, we can get LESS implemented in OpenSimulator if we really like it... the point is.. we need to like it :D
[11:23] Zha Ewry: he's ready to roll
[11:23] Zha Ewry: But, he actually needs to have code with two code bases
[11:23] Zha Ewry: and, honestly, if he provided a decent opensource code tree of the stuff
[11:23] Zha Ewry: I'd look at doing it in OpenSim, for fun
[11:23] Morgaine Dinova: Indeed
[11:24] Zha Ewry: the OGP proposals have three impls
[11:24] Zha Ewry: Mind you none of them are spec compliant, I think
[11:24] Zha Ewry: what's on VAAK is close
[11:24] Zha Ewry: PYOGP is... getting there
[11:24] Zha Ewry: I'm mising a few caps yet
[11:24] Meadhbh Oh: right. there are problems with the code on Vaak
[11:24] Meadhbh Oh: it needs to get cleaned up
[11:24] Meadhbh Oh: and we need to agree to a lot of things
[11:24] Zha Ewry: So, when Jon has that much code
[11:25] Zha Ewry: or *anyone* does
[11:25] Meadhbh Oh: (like content negotiation)
[11:25] Morgaine Dinova: And they're all kindof "lookalike" systems, so the interop is, well, not really as demonstrative as it might be.
[11:25] Zha Ewry: and they want to do a spec
[11:25] Zha Ewry: hey, come and dance
[11:25] Zha Ewry: well, all three were totally ground up impls
[11:26] Morgaine Dinova: But implementing the same model, so it's a corner of the problem space.
[11:26] Zha Ewry: (in spite of some people's complaining, the only thing OpenSIm reverse engineered was what goes on the wire)
[11:26] Zha Ewry: Which is totally within the IETF's remit. Get a solid corner, grow it out
[11:27] Zha Ewry: I'd *love* to see someone take OGP and add some different stuff to it, it would be ideal to validate the approach
[11:27] Zha Ewry: the IETF is very much about gradulism and incrementalism and evolution
[11:27] Zha Ewry: Which is why its specs tend to work, if they move at all
[11:27] Zha Ewry: and also why they often get oevrtaken
[11:28] Morgaine Dinova: "Grow it out" in the sense of make it more embracing of diversity, sure.
[11:28] Zha Ewry: /men ods
[11:28] Zha Ewry: exactly
[11:28] Meadhbh Oh: and for the record... i was offering my support for the croquet peeps, if they want to use LESS _and_ they want to have interop with SL
[11:28] Zha Ewry: chuckles
[11:28] Zha Ewry: Who replaced Meadhbh with a lawyer?
[11:29] Meadhbh Oh: @morgaine. for $diet{y|ies} sake. if you don't like OGP, write your own spec, get people to implement it, and then bring it before the IETF
[11:29] Zha Ewry: Here's the big thing, we can make the tent as big as we like
[11:29] Meadhbh Oh: if you do that, i'll help you with the SL interop
[11:29] Zha Ewry: But, the only stuff which is going to get real IETF blessing is stuff which both is in the tent, and has mutiple code impls
[11:30] Zha Ewry: I will, strongly agrue for the charter to be big enough to pick up more stuff, as it hits that level
[11:30] Zha Ewry: But, I will strongly resiist waiting on useful stuff, while waiting
[11:30] Meadhbh Oh: @zha. i'm for that, but it won't get approved if it's too vague
[11:30] Meadhbh Oh: (the charter that is)
[11:30] Morgaine Dinova: Talking about growing it out to make it more embracing of diversity ... how are we going to do that with LLSD (let's assume it's accepted)? It's been ripped to shreds on technical grounds, so at the very least it will need a lot of patching. But as a Linden spec, it's quite hardwired, and there is severe resistance to it changing, judging by the list. How might that proceed?
[11:30] Meadhbh Oh: it's supposed to have milestones and objectives
[11:30] Zha Ewry: We'll have enough concrete stuff to avoid that
[11:30] Meadhbh Oh: with the idea that you can recharter later
[11:31] Zha Ewry: Thus the charter will say' Her'es the space, and the milestones are focused on what's here, today, with mutiple code"
[11:31] Zha Ewry: and.. here are some work items, which we hope will lead to inclusion in the charter as milestones
[11:31] Meadhbh Oh: ohoh... right... morgaine... what you may not have known... there is a REQUIREMENT that charters be highly focused. and in fact, there were a couple of charter drafts that went back and forth between zha, i, and lisa
[11:32] Meadhbh Oh: some of the input was... "make your objectives more concrete"
[11:32] Teravus Ousley: You could also say that OGP.. is OGP in name.. in that it means (Open Grid Protocol).. Though.. that implies that there is definately a 'Grid'
[11:32] Zha Ewry: The balancing act, to my taste is "broad overall scope, focused objecteives"
[11:32] Meadhbh Oh: which is why i plotzed when JHurliman and you added "general consensus" as an objective
[11:32] Meadhbh Oh: cause in IETF charterland, objectives and milestones are supposed to be concrete
[11:32] Zha Ewry: Because that lets us say 'Here is the fertile ground for new stuff to add tot he charter"
[11:33] Zha Ewry: but.. keeps us on objectives and milestones which we can actually take to bakeoff
[11:33] Zha Ewry: and.. I'm ttoally happy to carry a milestone which is
[11:33] Zha Ewry: "New candidates for milesteons and bakeoffs"
[11:33] Meadhbh Oh: i would argue that virtual worlds hat are implemented on only one machine are not the perview of the IETF
[11:33] Zha Ewry: and let Jon, and Co. bring stuff there, when they have two impls to talk about
[11:33] Meadhbh Oh: @zha.. we need to talk to lisa about that
[11:33] Zha Ewry: Thus a BOF
[11:34] Zha Ewry: I'm not suggesting we delete any of the concrete milestones
[11:34] Meadhbh Oh: ack
[11:34] Meadhbh Oh: gotta run
[11:34] Teravus Ousley: I dunno, seems like a 'hurdle'
[11:34] Zha Ewry: I'm willing to entertain additional concrete work products
[11:35] Teravus Ousley: If it's good enough.. and we want it.. we might have to jump the hurdle for him.
[11:35] Zha Ewry: Jon's stuff?
[11:35] Teravus Ousley: .. only if we want it though :D
[11:35] Zha Ewry: Aye.
[11:35] Zha Ewry: If Jon emits a major language binding of LESS
[11:35] Zha Ewry: I'd be very tempted to see how hard it would be to build a co-simulation modeul for OpenSim
[11:36] Zha Ewry: I won't do it if they won't at least code up a complete working example
[11:36] Zha Ewry: but, hand me the 1500 lines of C++ to test against, and it's just not that bloody big
[11:44] Teravus Ousley: ok, well.. 4 drafts.. 30 minutes per drafts.. I guess.
http://slurl.com/secondlife/Tenera/208/83/70 (Starts in 15 minutes)
http://slurl.com/secondlife/Tenera/208/83/70 (Starts in 10 minutes)
http://slurl.com/secondlife/Tenera/208/83/70 (Starts in 5 minutes)
[11:59] Zha Ewry: Off to actually code for a few minutes
http://slurl.com/secondlife/Tenera/208/83/70 (Starts now)
http://slurl.com/secondlife/Tenera/208/83/70 (Started 5 minutes ago)
[12:10] Saijanai Kuhn: OK making transcript
[12:11] Morgaine Dinova: It makes fun reading. Just been going over it.
Saijanai

AW Groupies/Chat Logs/AWGroupies-2009-03-03

Navigation menu

Search