and solutions could be applied to collaborative spaces such as Wikis, Micro blogging
streams and collaborative desktop and application environments. "
[9:35] Latha Serevi: Hey, before we dive fully into the MMOX kool-aid pool for today ... remember how LL was going to announce some decisions and stuff early in '09? Is everything interop subsumed by MMOX (I think not), or just overwhelmed by it for am onth?
[9:37] Latha Serevi: It makes sense that we would put our limited energies into MMOX until the IETF meting - I just wanted to get a little perspective before immersing aghain.
[9:37] Zha Ewry: And, so far, the answe would appear to be yes, its ongoing
[9:39] Zha Ewry: One thing which Linden (an at times, many of us) need to do is stop thinking about this all as just Linden centric, but sure. Linden's saud, "The UDP gorp, never gets standardized"
[9:40] Morgaine Dinova: I am wondering if we are in "MMOX role" atm, or if we've got our SL/LL/OGP hats on
[9:40] Tammy Nowotny: one issue with interoperability is that there are different ways to model space in VWs... SL uses a 3-d world which is biunced by Z=0 and is otherwise infinite (unril you run out of integers at least): There.com uses, if I understood what they said at a recemt Metanomcis, a speherical world which is the size of the Earth. Google Earth is not a world per se but it is spherical and they are building Google Moon, Google Mars, etc.
[9:42] Gareth Ellison: i meant object positions etc are essential
[9:42] Tammy Nowotny: that is just a concent I thought I shd bring up: the basic geometry may differ from world to world, and we shdn't limit the standard to a SL-like world
[9:42] Gareth Ellison: UDP is just part of the implementation
[9:42] Morgaine Dinova: Zha: this is why I'm asking the question about hats. There appears to be a conflict here
[9:42] Saijanai Kuhn: well we covered that a bit by assuming non-cartesian worlds as being possible
[9:43] Morgaine Dinova: Therefore I'm assuming we have our non-SL hats on here
[9:43] Morgaine Dinova: Of course we could review it from an SL-protective perspective too,m but that';s a bit regressive
[9:43] FWord Utorid: what's the topic today anyway?
[9:43] Gareth Ellison: what i'm basically saying is, if you take OGP and define a nice basic "this is how you move objects" you have more of a generic VW standard
[9:43] Dedric Mauriac: can't you add a flag that says what type of coordinate system the object or world uses?
[9:43] Gareth Ellison: fword - Zha made a post to the mmox list
[9:43] Gareth Ellison: read the scrollback for an archive link
[9:47] Gareth Ellison: morgaine - your comment about 2D makes me think
[9:47] Zha Ewry: and stareted to try and define what we might be talkig about by "interoperaing"
[9:47] Morgaine Dinova: Which is what a BoF should be, so that's good prep work
[9:47] Gareth Ellison: we have MIME types on objects to rescue us
[9:47] Dedric Mauriac: if an object is made of many vecords - are we defining the coordinate type for each vector? i'm concenred over redundancy in "ccordinate-system/float/float/float" why can't the object itself identify the coordinate system
[9:47] Zha Ewry: and.. hopeflly reminded people firmly
[9:51] Gareth Ellison: i also like the approach of streaming events/movements etc with MIME types
[9:52] Morgaine Dinova: Prims are great ... even mesh-based worlds have prims, though they don't call them that ... they hand-craft mesh objects and re-use them. Prims are efficient, so good for the standard, even for mesh-based worlds. But of course, they should be OPTIONAL.
[9:52] Gareth Ellison: "here comes a stream of movements in (float,float,float) offsets from current pos"
[9:52] Saijanai Kuhn: I think this is where input from Croquet developers would be good
[9:52] Zha Ewry: I actually think, that one thing that's come up and been really postiive is that "streams" and in paritcular "Non immeersive strems" and how to share and sync them is important
[9:52] Gareth Ellison: morgaine - yeah, i think we agree that prims can be just one of many types
[9:53] Zha Ewry: Croquet, offers an interesting challange, that, they build up from t-objects which is a very non portable constryuct
[9:53] Yann Dufaux: accepted your inventory offer.
[9:53] Gareth Ellison: zha - streams, that's one i have running code for, and has 10 billion different other possible methods
[9:53] Morgaine Dinova: We all have monitors that can display things of more than 3 dimensions. After all, our monitors can't display 3 dimensions either --- it's a projection.
[9:53] Gareth Ellison: i think the question there is "what transport?" and "what serialisation?"
[9:57] Zha Ewry: For the nonce, I am seperating out "non merged streamed content" from "Merged streamed content with scenegraph elements in it" in my head
[10:02] Zha Ewry: as opposed to Binary streams of codec fodder
[10:02] Milele Demozay: There's also DTLS for secure datagram traffic but it's pretty low-level and the higher-level transports ought to be examined first.
[10:02] Zha Ewry: (mind you, if the IMVU suff doesn't move as binary bits, I'm sort of umm. Disblieveing"
[10:06] Morgaine Dinova: The SIP if used would be something like the control stream in FTP though, with the actual data being handled over SCTP. ie SIP only does inittiation/teardown
[10:06] Zha Ewry: and.. I'm also lookingfor us to re-ground our discussoin a bit in "What code will we actually embody this in in the next two years"
[10:06] Gareth Ellison: zha - i'll have to post a more detailed response on the list
[10:06] Zha Ewry: Thanks Gareth, I look forward to it
[10:06] Gareth Ellison: but my general thoughts are that we need the MIME types for objects, and we need streams
[10:06] Zha Ewry: I'm not expecting real time feedback here, more.. pushign people to do exactly that
[10:06] Gareth Ellison: what code - something in python or C
[10:07] Zha Ewry: and ask any questoins that woudl hepl witht hat
[10:08] Gareth Ellison: well, i'd like to publically apologise for my badly thought out permissions proposal
[10:08] Zha Ewry: I'd argue that the other implicit one is "Keep ocean boiling to a minium"
[10:09] Zha Ewry: Its easy to get all excited about some of the really big picture stuff
[10:09] Saijanai Kuhn: Gareth the permissions thing is called "the third rail" by LIndens, so not to worry
[10:09] Morgaine Dinova: Zha: since we have to implement this, I don't think there's a danger there :-)
[10:09] Gareth Ellison: i was thinking in the context of "i want to move stuff around without pestering copyright holders, and it's not going to enforce anything", but it turned into an ugly DRM thing
[10:14] Milele Demozay: We frequently have chairs who are employed by companies who have an interest in the technology
[10:14] Infinity Linden: hey folks... sorry for being late... meetings
[10:14] Morgaine Dinova: I think it's important that either Infinity is SEEN to be agnostic, or that when she presents her Linden case then she says that she is doing so.
[10:14] Zha Ewry: the IETF assumes that the chair cts as an individual,
[10:15] Milele Demozay: You have to remember that an engineer's point of view is deeply influenced by what they know well, but the way to deal with that is not to yell "BIAS" but to educate them on alternative POVs.
[10:16] Gareth Ellison: my god, you don't want to know how much me and a bunch of others were heavily SL-influenced with a totally different project
[10:16] Infinity Linden: also keep in mind that there are reasons we did things the way we did... i've been "in the middle of it all" so i have an intuitive understanding of the technology decisions we've made
[10:18] Zha Ewry: Which is that people like Linden (and possibly OpenSim implementors) are constrained by exiesting code
[10:18] Zha Ewry: It would be really good, if we can get superset solutions
[10:18] Zha Ewry: not ones which say "rip up huge amounts of code"
[10:18] Infinity Linden: @gareth... i agree... but back then i was just a beta tester and many of the peeps who were implementing the code came from a games background who said ... "oh my $DIETY! i must use UDP"
[10:19] Zha Ewry: and then promptly re-did the TCP layer badly
[10:19] Infinity Linden: and i should also point out that there are some very selfish reaons for using UDP in an environment where TCP is driving traffic to the network rate
[10:20] Gareth Ellison: i can see possibly overhead, but such overhead can't be THAT expensive
[10:20] Morgaine Dinova: Zha: we're not defining an architecture though, but only a protocol and interface --- so really existing worlds aren't constrained by their existing code too much --- they don't NEED to handle everything, but just go for limited exchange initially.
[10:20] Infinity Linden: (that is... we can generally use UDP to cause the comparatively less-flexible windows TCP stack to begin traffic back off earlier
[10:21] Gareth Ellison: it smells like horribly premature optimisation
[10:21] Zha Ewry: if we do stuff that makes it really hard to migrate from the current code bases, we're likely to see people just throw up thier hands in disgyt
[10:22] Infinity Linden: but i should probably point out.. we're interested in a highly constrained version of "interop"
[10:23] Gareth Ellison: that's the approach i used to drag my avatar out of SL with group chat intact
[10:23] Morgaine Dinova: Well, it's not about sims --- those are implementation details. We define interop endpoints, and how an architecture maps those is its business.
[10:23] Zha Ewry: since anything epressed as a webservices endpoint is easy to proxy
[10:23] Gareth Ellison: and to TP into arbitary osgrid.org regions via OGP before OGP was deployed there
[10:23] Infinity Linden: LL has no interest in making a general purpose client that allows users to connect to SL and WoW and qwaq and OLIVE based systems
[10:23] Zha Ewry: Whichi isn't interop, so much as unification
[10:23] Gareth Ellison: zha - i meant a proxy of the "converting protocols" type
[10:23] Infinity Linden: gareth... please... shall we not talk about the way you abused the ToS?
[10:23] Morgaine Dinova: Yep, the endpoints are service endpoints. Not sims in the spec, although of course if a company chooses to expose internal servers that's its business.
[10:27] Zha Ewry: But, people should be able to build much smaller clients and have it all work
[10:27] Gareth Ellison: basically, each single world needs an extra listener for the new protocol we define
[10:28] blotto Epsilon: Infinity, re "highly constrained version of interop" -- can you say what kind of interop LL does want to have?
[10:28] Zha Ewry: see my comments about "growing out interoperable puddles"
[10:28] Gareth Ellison: that extra listener can be in the same servers (existing servers), or can be with the proxy method
[10:29] Infinity Linden: we like interop, defined by the community, that allows us to implement a protocol that is usable by client software other than our own
[10:29] Zha Ewry: I expect someone will build a "swiss army client"
[10:29] Rex Cronon: if is a very light client, u could possible have multiple instances of it connected at same time to different VW
[10:29] Infinity Linden: and that client software SHOULD be able to connect to other worlds
[10:31] Zha Ewry: The place where it doesn't work, is when the granulaity gets wonky
[10:31] Tammy Nowotny: soeme of the old mainland sims are 90% plwyood boxes already. LOL.
[10:31] Morgaine Dinova: If LL's position is that they will only support a "highly constrained version of interop", then it is even more important that we hear when LL personnel are speaking on behalf of LL or acting as even-handed co-chairs.
[10:31] Gareth Ellison: infinity - the prims vs meshses are seperate MIME types :)
[10:32] Infinity Linden: Zero and I are in the process of pelting our engineers with rotten vegetables from the "separation of concerns" farm
[10:36] Gareth Ellison: you also have IE which screwed up PNG transparency for a long time
[10:36] Gareth Ellison: but the point is that "it still works"
[10:37] Gareth Ellison: you don't need to support every possible kind of image to make a web browser
[10:37] Zha Ewry: with proper mime types and all in the right places
[10:37] Infinity Linden: i have an apple dashboard widget that lets me god shout in various regions and do various grid monkey things. it was annoying to develop
[10:37] Morgaine Dinova: Our systems will evolve, I don;t think we should be dismayed that there will be the equivalent to the web's "broken images" in the early stages.
[10:37] Infinity Linden: it would be very nice if we coul dhave a system that allowed the client to register what things it's interested in
[10:37] Tammy Nowotny: but you do need to support the formats peoole actualy use... especially since you can't count on them to add the appropraute metadata
[10:38] Infinity Linden: so the "lynx viewer" would maybe display some gross geometry stats
[10:38] Asterion Coen: zah when a monetary is going up, an other will go down. the l$ use to be about stable , so u can buy some usd, resell in GPB, or euro, depending of the better rates without some risky operations :)
[10:38] Zha Ewry: just pointing out that, we want to be able to do any encomic activity in a currency agnostic way
[10:38] Infinity Linden: (which.. btw.. is why we do our auth the way we do... after auth, you request which capabilitues you want to use)
[10:40] Morgaine Dinova: Infinity: inevitable though, that's what interop will be about, until the days there's just one cohesive metavers, is is probably never
[10:40] Elric Ember: Fail gracefully for granularity mismatches by substituting a default rendering in the scene graph or apply a transform to the input (ala xslt mapping, etc) providing a rendering that makes sense in the receiver's context.
[10:43] Zha Ewry: "Given all this, how do we tie this to specific specs like OGP, LLSD, ad X.509?"
[10:43] Morgaine Dinova: In any case, in the early years we're going to have tons of "broken objects" --- every time that someone TPs somewhere wearing an item from an SL creator that doesn't want interop.
[10:44] Yann Dufaux: well, i keep firefox and i wait for firefox 4 :)
[10:44] Infinity Linden: morgaine... can you add context to that incendiary comment?
[10:44] Infinity Linden: (honestly... it can be taken a lot of ways...)
[10:45] Milele Demozay: Yes Zha, and old-time IETFers are going to have a pretty hard time looking at the specs already published and see how that's necessary for MMOX
[10:45] Zha Ewry: I take it as happily out of scope. We'll define mechanisms which permit people to deploy various IPR regiemes and step back and watch evolution happen
[10:45] Milele Demozay: BTW, IETF meeting etiquette/rules are that attendees of the physical meetings read all the docs listedin the agenda beforehand.
[10:45] Morgaine Dinova: Infi: it's not incendiary, just factual. I assume you're going to respect creators wishes, so after TP, people wearing things without interop perms will not have them.
[10:46] Infinity Linden: well.. we're NOT going to use ASN.1
[10:46] Gareth Ellison: like, to be nuts some kind of "SL-UDP-packets-streamed-over-http" MIME type
[10:46] Zha Ewry: got told by an old IETF hand "Its not enough to pass your eyeballs over it"
[10:46] Zha Ewry: "You actually are supposed to try and make sense of it"
[10:46] Infinity Linden: and XDR doesn't have the types we need
[10:47] Infinity Linden: and both have semantics that are show-stoppers for us
[10:47] Zha Ewry: I think, Infinity, we need some use cases and some objective statements which say that
[10:47] Milele Demozay: The cutoff for submitting new -00 drafts is March 2, in case anybody wants to add to the list of what we might expect IETF oldtimers to read to prepare themselves.
[10:47] Zha Ewry: and probably some explanatoin about why XML with DTD and SCHEMA isn't sufficient
[10:47] Milele Demozay: In the long run there's no difference in the way we treat people who come from VW experience and people who come from IETF protocol design experience, but at the BOF there will be a deep gulf
[10:49] Rex Cronon: considering that in sl is possible to copy textures, 3d objects, is very unlikely that anything from WOW will get here any time soon
[10:49] Infinity Linden: i would be happy to chat with her if she wants to go for a new XML serialization
[10:49] Morgaine Dinova: Lisa: how do updates to drafts beyond -00 get put up? We currently have the problem that we've make 4 iterations of the original charter draft, yet the IETF wiki still only shows the original, so all our work on updates us currently not visible.
[10:49] Zha Ewry: Well, you can capture all of those in Wow too
[10:49] Infinity Linden: these things aren't RFCs at the moment
[10:50] Infinity Linden: @morgaine.. there's an ID draft submission tool
[10:50] Gareth Ellison: not much of a wiki when nobody can edit it
[10:50] FWord Utorid: the thing about all of this that is happy tricky and people forget while we go for standards and stuff.... the guy who wants to sell his stuff benefitted from the fact that the protocol was closed for a while
[10:50] Zha Ewry: Now, now, Gareth, they all end up lik that anyway
[10:50] FWord Utorid: when it becomes common knowledge how it all works and so on there's no commodity
[10:51] FWord Utorid: and Infinity's L$ checks bounce :P
[10:51] Zha Ewry: lurves wikis,but they are such bogs of bit rot
[10:51] Infinity Linden: i believe that if you made a new draft, you would be a new author so it would be something like draft-<your actual last name>-vwsd-00.txt
[10:51] Milele Demozay: Yeah we're not very good at using wikis in IETF, sorry
[10:51] Zha Ewry: How do we mark "rough consensus drafts" ?
[10:52] Milele Demozay: Drafts that a WG has accepted are draft-ietf-<wgname>-subject
[10:52] Tammy Nowotny: although some people will alwasy opt to buy instead of make... making takes time and knowledge, and some p[eople have limited amounts of those things
[10:52] Infinity Linden: @FWord. people don't benefit from protocols being closed
[10:52] Zha Ewry: I'd like to be able to say "This is one we'd like to propose as one that has a otugh consensus"
[10:52] Milele Demozay: But that doesn't mean that the WG has come to consensus yet on the content
[10:54] FWord Utorid: @Infinity, of course people benefit from protocols being closed, or nobody would have shit bricks about things like copybot
[10:54] Milele Demozay: Zha: you can put links to individual drafts in the charter, does that help?
[10:54] Morgaine Dinova: Lisa: currently we've been using a pass-the-baton approach, so people have been improving each other's previous versions. Should we ask a co-chair, or the AD, or what, to post a new version to the wiki? Should co-chairs we asking "Is this OK folks for a new revision?" Or what? How does it work?
[10:54] Infinity Linden: but... we have a closed protocol at the moment and we also have copybot
[10:55] Milele Demozay: It's OK to have it somewhere else as long as it's periodically copied to the list
[10:55] Infinity Linden: you don't get to redefine the draft charter becuase you don't want OGP published
[10:55] FWord Utorid: don't sweat it, I am just here for target practice.
[10:55] Tammy Nowotny: open source protocols are usually more secure, in fact
[10:55] Infinity Linden: it's something that must be a discussion in the community as a whole
[10:55] Infinity Linden: if we move LLSD to being an informational RFC
[10:55] FWord Utorid: just umm... don't go into space without a helmet
[10:55] Infinity Linden: then the OGP RFCs that follow cannot be standards
[10:55] Latha Serevi: I also find the volume of discussion a bit overwhelming. So, although the MMOX mailing list should probably be the "on the record" gold standard, I wish some kind soul would maintain a "executive summary" stream or blog, and that we had a set of scratch pad wiki pages (perhaps Groupies/MMOX, or is that too SL?)
[10:56] Infinity Linden: (as they depend on LLSD as a normative reference)
[10:56] Zha Ewry: they have to be informatoinal as well
[10:56] Zha Ewry: becuasee you can't have an normative depdn on an informational
[10:56] Morgaine Dinova: Lisa: yes, the revisions are in the mailing list. The trouble is, some people are continually posting links to the original, thus totally ignoring the process that lead to the revisions.
[10:58] Zha Ewry: what we effectively have is 3 or 4 seperate wordings being proposed
[10:58] Milele Demozay: Ultimately the people chairing the BOF get to decide what version to present at that meeting
[10:58] Infinity Linden: Miele... certainly morgaine doesn't have to submit a new draft charter for a new group, right?
[10:58] Milele Demozay: I asked Infinity and Barry Leiba (very experienced IETF colleague) to chair the BOF
[10:58] Infinity Linden: ugh. i think this debate is going to go on for a while
[10:58] Infinity Linden: and i have to step away from being a linden
[10:59] Zha Ewry: (I'm going to get to be the pesky person who's not biased) to throw tomatos in SF
[10:59] Milele Demozay: Doesn't have to? No, but can -- it's considered polite to suggest new wording, compared to just complaining
[10:59] Morgaine Dinova: The way I *expected* it to work, is that one or other co-chair would see a consensus being formed, and then post up a new draft. But it hasn't happened.
[10:59] Infinity Linden: so my objective is to get morgaine's input so she doesn't say "she's being partisian"
[11:02] Morgaine Dinova: No apologies needed Infi. :-) I just want to see everybody's work getting incorporated (including mine). I don't want the BoF to start with a charter that *unanimously* on the list (including Infinity), everyone acknowledges is heavily biased towards the LL/IBM view.
[11:02] Zha Ewry: I defy you to find anything concrete enough in the screed to acuse of being biased ;-)
[11:02] Tammy Nowotny: some places are half an hour off from the neighboring time zones, Goldie... e.g., Newfoundland and the 2 central states in Australia.
[11:03] Milele Demozay: The Charter revisions will go on well after this first BOF, I am quite certain
[11:03] Tammy Nowotny: I didn;'t think you were too parysan Infinity
[11:03] Infinity Linden: right. but i think when we proposed it, we really were thinking... "hey... here's protocol (OGP) that's being used in a couple of places.. sure would be nice to have it standardized."
[11:03] Milele Demozay: of course the more progress we can make before, the better
[11:03] Zha Ewry: And.. seriously, what I want is to get a broad superset of people's needs, which includes what Linden and various players are going to do
[11:03] Infinity Linden: we weren't really thinking the discussion would turn to "how do i use my SL credential to log into WoW"
[11:04] Zha Ewry: No, but you probably have to be asking 'How to I do presence between Wow, SL, Olive and QWAQ?"
[11:04] Tammy Nowotny: I see two typos in the names on the page..