[9:36] Zha Ewry: chuckles, I stick to the SL timezones, to avoid complete confusion
[9:37] Zha Ewry: So.. Since I know what 90% of people have been talking about on the IM stream.. I'mgoing to start with a quick positioning observation
[9:37] Susan Tsuki: so, every three months they want to exchange concrete (presumably that means working) proposals and hope that someone in the AWG implements them in an OpenSomething so they can get IETF interoperability approval?
[9:38] Zha Ewry: I don't speak for Linden, I don't know what Zero meant by his note, I'm not about to spend an hour speculating, when he'll be on in 3 hours
[9:38] Lor Gynoid: Concret working proposals instead of those carbed into tablets of stone? :)
[9:39] Susan Tsuki: and if they get concrete, working proposals that are already implemented in an OpenSomething, then they might try to do those too for more IETF draft promotions
[9:40] Zha Ewry: I'd look at it as more, "We're not interested in design input into our code internal, bring concrete proposals at IETF draft input ready level to the table"
[9:40] Susan Tsuki: the quarterly thing is confusing, all the IETF stuff is done in queues. Maybe they mean they expect average turnaround times of 4.5 months for reviews
[9:40] Pixel Gausman: Susan: i'd like to see GridInfo as an IETF proposal
[9:40] FWord Utorid: ok. they're gonna write specifications and form a committee to talk to other committees.
[9:45] Tao Takashi: Zha: What are you working on (Sorry, a bit out of the loop)?
[9:45] Goldie Katsu: So on first glance, unless the IETF stuff has a bigger/faster effect than I think it seems to me this forces a faster drift/split between Open Sim and SL platforms
[9:46] Susan Tsuki: is there an IETF Working Group which entertains SL protocols yet?
[9:48] Zha Ewry: The big advantage, is it ought to give us a place to push on some of the ideas we're intersted in, ina OGP context, without needing Linden to deploy stuff
[9:48] Susan Tsuki: in that case, you must title the draft filename: draft-yourlastname-projectid-specid-01.txt
[9:48] Tao Takashi: unfortunately I am drowned in work. Otherwise I wanted to try to hook up some OpenSocial/Oauth experiments with OpenSim
[9:49] Susan Tsuki: you might want to settle on a projectid so it's easy to find all the SL-related drafts in the database. slawg?
[9:49] Zha Ewry: I think Infinity has been trying to keep it at the Massivlye Multipparty space
[9:51] Pixel Gausman: Morgaine: and that is evil why?
[9:51] Susan Tsuki: "Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet-Drafts. " that means you have to submit drafts before they will recognize a working group, I'm pretty sure
[9:52] Susan Tsuki: right, if they submit stuff to the IETF, it means they promise not to sue you with submarine patents if you implement it or copy it
[9:52] Teravus Ousley: So far, in my experience, corporations and business minded, tend to write standards that are.. the easiest for them to implement.. not the best
[9:52] Pixel Gausman: Dahlia: depends on how wide the interests are in the group, and how good existing draft proposals are
[9:53] Zha Ewry: When there is a clear draft, ti can be fairly swift
[9:53] Zha Ewry: (If there isn't a huge "NO, not that approach" debate
[9:54] Pixel Gausman: i think that the standardization process will still be slow going.
[9:55] Zha Ewry: I think beyond LLDL, and few bits, its going to be a bit of a slog
[9:55] FWord Utorid: standardization processes always are. and, they are later embraced, and extended!
[9:55] Morgaine Dinova: Well there's nothing to standardize yet, after 15 months. So yeah, it'll be slow.
[9:55] Pixel Gausman: LLSD is a pain point for the OpenSim community
[9:55] Zha Ewry: On the other hand, I'd like to have a place where we can say "If you build it to this"
[9:55] Zha Ewry: there is a chance it will be usable
[9:56] FWord Utorid: idk that standards bodies are required to legitimize what has already 'happened', but it's a logical step in lethargy to involve more people
[9:56] Dahlia Trimble: ya who needs consciousness anyway
[9:57] Latha Serevi: What kind of layers might we see, or want to see, in the IETF drafts? e.g. authentication, messaging, 3-d prims, scripting
[9:57] Morgaine Dinova: I read "What kind of lawyers ..." ;-)
[9:59] Tao Takashi: another question I would like to ask is if everything should be based on the stuff Linden Lab invented even if there are already standards out there doing the same
[10:00] FWord Utorid: it is not my way, zha. it is the way people are going because it's available and not a committee
[10:00] Rex Cronon: what exactly did ll "inveted"?
[10:00] Dahlia Trimble: There is going to be a hypergrid meeting on Friday I believe...
[10:00] Teravus Ousley: I would also like to point out that I have pledged to implement OGP drafts as they are released. So, OpenSimulator will not 'drop' OGP support there.
[10:00] Morgaine Dinova: What would an OGP-Hypergrid hybrid look like? Because LL are going to totally block any attempt at going back on their work so far.
[10:01] Lor Gynoid: is concerned some quite big chunks of "new stuff" will need to be invented...
[10:01] FWord Utorid: I don't think anyone should assume that you will be teleporting back and forth with full everything evar
[10:01] Zha Ewry: And, I'm pretty much going to keep psuhign the OGP stfuff as well
[10:01] Tao Takashi: like LLSD; like capabilities and probably much more
[10:01] Latha Serevi: Thanks for your specific answer on layers, Zha (LLDL, Caps, X.509 PKIX stuff) and Pixel's addition of "3d Object Formats". Is there a URL that discusses this topic that anyone knows of?
[10:01] Teravus Ousley: that being said, hypergrid is cool in many ways.
[10:02] Susan Tsuki: the IETF has far and away the very best record of standards adoption in the information processing world, save for perhaps a dew Federal Information Processing Standards
[10:02] Zha Ewry: I think, realistically, from some grids to some grids, there will be al ot of stuff which moves
[10:02] Zha Ewry: Just as from others, the amount of movement will be close to zero
[10:02] Tao Takashi: sure there needs to be new stuff invented but the whole social networking layer of Second Life could be done by using social networking standards IMHO
[10:02] Teravus Ousley: I don't think there is any issue with hypergridding into one region and OGP-ing out
[10:02] Morgaine Dinova: Zha: do you mean that you too will block any Hypergrid-type functionality?
[10:02] Pixel Gausman: gets called away and waves byebye to AWG
[10:05] FWord Utorid: there is the occasional turbo mode rocket that gravity pulls back to the ground in flames
[10:05] Morgaine Dinova: ANd thinking a little, it might emerge that supporting a view that isn't fully corp-centric might actually help.
[10:06] Zha Ewry: So, expect some work on what happens when you do both OGP and Hypergid teleports
[10:06] Zha Ewry: and what a good sbstraction for locating remoted UUIDs is
[10:06] Tao Takashi: you know, even if corporates might dominate something now it does not need to stay that way.. we are still in early VW times and lots can happen
[10:06] Tao Takashi: and actually the most promising things even in VW world are already happening in an open source basis
[10:07] FWord Utorid: tao: it will stay that way unless you write it all and make it open source and non-extendable and hire corporate lawyers to defend your IP
[10:07] Zha Ewry: I'm expecting to be able to teleport between various regoins usinf both sceme
[10:07] Tao Takashi: and OpenSim could probably also support 10 different connection methods ;-)
[10:07] Zha Ewry: and over time, sort out the various authentication and policy issues as we move alnog
[10:07] Zha Ewry: I've said, for a while (go find my comments at Vw2008, in LA)
[10:07] FWord Utorid: and in order to defend your IP and get tax breaks and expand.... you will need to form a business entity
[10:08] Susan Tsuki: the IETF has been less susceptible to corporate pressure in the past, compared to, for example, the IEEE, because the IETF motto is "rouch consensus and working code" and their process requires actual interoperable implmentations before they will promote anything to a Request for Comments, which is the usual stage around which internet standards coalesce. Whereas with the IEEE your company can submit an entire closed software system, and as long as the full color glossy "documentation" brochures (usually "white papers" plus a graphic designer) as submitted along with allowances for a sufficient number of deductible junkets, the IEEE and similar will essentially sell a stamp of approval. The IETF has for the most part withstood such pressures
[10:08] Zha Ewry: that you need both OpenSource and Standards to drive the full range ofinnovation we want
[10:08] Susan Tsuki: "rough consensus and working code"*
[10:09] Susan Tsuki: the IETF won't give you a stamp of approval unless competitors have actual interoperable implementations, unlike practically all the other standards bodies
[10:09] Zha Ewry: and working code to show interop
[10:09] Morgaine Dinova: We're discussing this in a single group here, but it rather looks like we'll all be competing against each other at IETF, going by Zha's "I'm paid by IBM" position.
[10:09] Zha Ewry: Not 100%, not a deployed product
[10:09] FWord Utorid: anyway, i don't know why people fear corporations. it's not like the guy in the basement is some sort of a hero
[10:10] Saijanai Kuhn: story of janitor who once entered a room full of Tmers at a high school and went "OMG they're all dead!!!"
[10:10] Saijanai Kuhn: he ran out of the room when everyone started laughing
[10:10] Zha Ewry: I intend to build good, sane code, with good sane spec, and I expect I will be guided by the people who pay me
[10:10] FWord Utorid: i think it's a 'finished' product that everyone is afraid of, because then there would be no need to rewrite it all
[10:11] Tao Takashi: show me a "finished" product please! ;-)
[10:11] Saijanai Kuhn: I'm learning iphone programming to port pyogp to ObjC on an iPhone. don't confuse the mess of the GPL viewer with a sanely designed OGP viewer
[10:11] Tao Takashi: I guess Second Life is one, right? ;-)
[10:11] FWord Utorid: tao: windows 3.11 for workgroups is finished
[10:12] Zha Ewry: I'm fairly indepdnent, becauee I sit in the Research division, but.. being paid does focus the mind. I've been pretty clear on that since day one of my involvement in the space
[10:12] Latha Serevi: Morgaine, you're sounding so cranky today! Isn't the agenda here to discuss the different ways that we love Zha?
[10:12] Morgaine Dinova: Zha: the trouble is Zha, those specs will inevitably be "sane" only for a strictly controlled, centralized grid that provides the kind of security IBM needs, which is a rather different proposition to what a "grid of the people" needs.
[10:12] Susan Tsuki: nobody is asking for a finished product before a spec, but that's okay too. But the internet drafts need to be submitted to the IETF before they can do their testing of them, which is known as a "bake off"
[10:13] Rex Cronon: how can u now that zha? do u happen to see the future:)
[10:13] FWord Utorid: notes that there's nobody from the IETF at the AWG meeting, and where is zero?
[10:13] Morgaine Dinova: Latha: I'm not trying to be cranky, I'm just frightened as hell at Zha steamrollering the IBM view at the expense of Opensim's.
[10:13] Saijanai Kuhn: suspects the lindens are deliberately stayiikng away today
[10:13] Goldie Katsu: There will be multiple points of view in any IETF process.
[10:13] Latha Serevi: I've spoken before about how our OGP efforts are hampered by the fact that there's no strong "pole" community to counterbalance LL. I hope the IETF process involves at least one other clear voice. I wonder what community might produce that voice?
[10:13] FWord Utorid: I am not afraid of IBM. They made OS/2 Warp.
[10:14] Zha Ewry: I'm paid to keep a broad perspective and I don't think that a corproate only perepctiev is wise
[10:14] Zha Ewry: The nice thing about the IETF is it isn't on Linden's wiki, or with Linden as the sole editor
[10:14] Tao Takashi: does the move to the IETF actually mean really a big difference?
[10:15] Dahlia Trimble: is an IETF standard required for implementation of a protocol any?
[10:15] Teravus Ousley: "more flair..." zha, you're not wearing enough according to the chart under the water
[10:15] FWord Utorid: and they were talking about writing specifications last year
[10:15] Rex Cronon: but, will ll actually implement what IETF approves?
[10:15] Susan Tsuki: at this stage, if you want to benefit from the economies of scale that standardization can provide, and LL clearly made the decision that's what they want, you should just make sure you can find each other's drafts, so remember to announce them to the sldev mailing list and name them something consistent so anyone stumbling across one can find all the rest; I suggested the filename template draft-[YourLastName]-slawg-[YourSpecsTopic]-01.txt (the -01 part is incremented for revisions to -02.txt etc.)
[10:15] Zha Ewry: That is, of course, a questoin for Linden Labs
[10:15] Tao Takashi: Rex: I would even like LL to specify something different (simpler maybe) than they implement
[10:15] Latha Serevi: Rex, I don't care if LL implements what IETF approves, although they must undoubtedly give it a nod. I like how it might give the non-LL community a way to coordinate better.
[10:16] Zha Ewry: I expect the spec to be a cleaner superset of what Linden builds
[10:16] Susan Tsuki: the IETF will never approve anything unless nominal competitors have already implemented it, that's the beauty of their system
[10:16] Imaze Rhiano: well... hopefully we will get more structural approach to this process... like meeting agendas and achieve something faster through IETF...
[10:16] Zha Ewry: And.. I expect the mechanisms to be seperate form the policies
[10:16] Zha Ewry: That's where I think Morgaine gets what she wants
[10:16] Saijanai Kuhn: if you put them on the wiki, I'll fix up a category that can be stuck at the end so anyone can find them
[10:16] FWord Utorid: I expect in 2010 to have this same meeting
[10:16] Teravus Ousley: They're going to use OpenSimulator as that competetor... so that's not a 'protection'
[10:17] Morgaine Dinova: Well LL are going to ensure that what IETF endorses ends up being 100% what LL has put in, no departures. We've seen no flexibility from Zero whatsoever, after all.
[10:17] Zha Ewry: I'm not going to view OpenSim as sufficient, if somenoe asks me for a vote
[10:17] Zha Ewry: The nice thing about the IETF, is that Linden is a single vote in the process
[10:18] FWord Utorid: hmm... ibm says opensim sucks at not being sufficient. teravus, have everyone add sufficient to the mantis
[10:18] Zha Ewry: Not sufficient from the IETF interoperable example level
[10:18] Zha Ewry: Its too closely wedded to the current client
[10:19] Susan Tsuki: if OpenMetaverseFurryFork, Inc. and SimianBundleFromPodunk, Ltd. aren't happy that LL refused to do some better authentication than they did, then they can get together, make an interoperable implementation, and leapfrog LL's non-interoperable thing, putting pressure on LL to reconsider their approach
[10:19] FWord Utorid: zha: yeah, because they want the backend to be able to work without reinventing the wheel
[10:19] Tao Takashi: maybe this whole protocol should even be splitted into different sub-protocols which then can be standardized
[10:19] Tao Takashi: like the social networking aspect (might even be broader then than just VWs) and the 3D part
[10:19] Zha Ewry: I expect a series of RFCs, not noe
[10:20] Zha Ewry: Starting with the LLDL, caps and such
[10:20] Tao Takashi: replace it with JSON and OAauth :-)
[10:20] Zha Ewry: Authenticatoin, AD, and and Regoin Domain, all seperate
[10:20] Saijanai Kuhn: MPEG has several groups working on such things. They're mostly not nearly as far along as AWG/OGP IMHO
[10:20] Hermit Barber: Tao: It depends. If IBM has an interest which is clear, compelling and different from LLs, and if the Open Sim movement can at least put a hat on the desk which is actually independent, possibly with support from IBM or some other player, because the new LL's management direction appears to be all inclusive but only if there is no competition, then the IETF process could become very interesting and helpful. But at the end of the day, the IETF does not *make* anything happen, they won't add bodies, money or other resources to the project, and in fact because it is another area requiring lots of time,effort and attention to detail, it adds load to the very few people that actually do the work in any community project. Which might be why more proposals sink of apathy than are implemented.
[10:20] FWord Utorid: replace it all with CAPSLOCK and SHIFTKEY
[10:20] Zha Ewry: Content and extendisble low level prims
[10:21] Morgaine Dinova: If the spec can be made a superset of LL's proposals, I'm happy (in fact, that's necessary, since we all want compatibility with SL). But LL are going to fight tooth and nail against that, because after all, what they're doing here is trying to control the space, and a superset implies loss of control.
[10:21] Zha Ewry: The ITEF has always been a supserset style
[10:21] FWord Utorid: LL wants control of the IETF?
[10:21] Zha Ewry: If you look at what happens as specs move through the IETF process
[10:21] Tao Takashi: Hermit: right, to the process itself the IETF does not help (so if people don't make progress it won't come from the IETF) and it might add additional overhead
[10:22] Zha Ewry: Competitors bash in stuff that they need, and you end up with a superset, which is often braoder than any one implementatoin.. There tends to be a fun process of winnowing of fetures
[10:22] Susan Tsuki: Morgaine: do you think that Zero could over-rule bakeoff results if a bunch of developers had a cross-grid teleport protocol for avatars, their assets, and inventory, and LL didn't? It would be against the rules and there's a lot of institutional opposition to any party who tries to manipulate a bake off like that
[10:22] Zha Ewry: oh.. and of course, one should recognize that some substantial fraction of the IETF drafts never get implemented
[10:22] FWord Utorid: I think Teravus and the opensim developers destroy the IETF and all lag generated in this process. who needs standards when we have toys?
[10:22] Saijanai Kuhn: well, LL could simply not apply that to their internal grid...
[10:23] Zha Ewry: I'll make two quick observations, and wander off to chase other horrible things
[10:23] FWord Utorid: I have an idea, let's write SPECIFICATIONS for twelve years.
[10:23] Imaze Rhiano: who can afford to wait standards - when competitors are running away... SL might become technical fossil...
[10:23] Zha Ewry: 1) the IETF gets you a place to define the spec which isn't owned by Linden Lab. I think that's a net puls
[10:24] Hermit Barber: Working code is the best specificationn.
[10:25] Susan Tsuki: LL wouldn't have to apply it, but when they are trying to sell servers, or get the monied corporate users to use the main grid, other server vendors can say, "don't get locked in: you wouldn't want a browser which could only view one site blah blah"
[10:25] Tao Takashi: Sai: not as a spec, but it needs to be implemented
[10:26] Zha Ewry: "Here is code which claims to do it"
[10:26] Latha Serevi: I alwo like how, if the IETF process gets stuck, that doesn't prevent anybody from writing code. It's a place and method for coordingating, that we sure need right now.
[10:26] Zha Ewry: The nice thing is that if the code happens to put all the serliazed parameters in alphabetical order
[10:26] Dahlia Trimble: everyone who wants to put up walls around the garden now knows that all they have to do is deviate from the spec ;)
[10:27] Zha Ewry: You can say "No, the alpha order is a implementtaion quirk"
[10:28] Morgaine Dinova: The trouble with Zha's 1), 2), is that it misses something absolutely key: extensibility. You can't create working code for something that isn't yet known, and I see LL blocking moves for extensibility as unjustified.
[10:28] Zha Ewry: I intend to build an extensible spec
[10:28] Saijanai Kuhn: Morgaine, the non-LL AD does away with that issuye, I think
[10:29] Morgaine Dinova: Sai: it might, but only if the set of cap is not constrained to the features that LL wants only.
[10:29] Zha Ewry: What I expect, to happen in some spaces, is that the spec will be broader than what Linden deployes
[10:29] Zha Ewry: With a ITEF extensible spec, we can do that without Linden's owning the documents, thats' a net plus
[10:30] Saijanai Kuhn: Morgaine, there's nothing stoping someone from adding new features in the AD and something to handle them, even in the GPL client
[10:30] Morgaine Dinova: But there's more to extensibility than just the caps available. For example, if objects are to become hierarchical in the future (as everyone wants, in theory), then OGP itself must be able to handle trees of objects, not just bottom-level primitives.
[10:31] Hermit Barber: Zha: If the spec is for the line protocol, with a minimal required set to establish sessions and capabilities and the rest being designed as optional but requiring published, working plug ins to be accepted, with implementations being allowed to do whatever they like as long as the line protocol keeps working and the plug-ins are compatible, then who cares. That may well be the best approach to an area which is in the turmoil of early development if an early spec as framework is sought.
[10:31] Saijanai Kuhn: and... I dont' see this as an insurmountable issue
[10:33] Zha Ewry: That's somewhat less likely to fly with the IETF, Morgaine
[10:34] Rex Cronon: what is so bad about "balkinization"?
[10:34] Hermit Barber: Morgaine Looking at the spagghetti that is SL, and the emergent corporate desire for ownership, I can understand why. But IBM and OpenSim and potential users all have other interests.
[10:34] Saijanai Kuhn: a certain tension there, of course, Morgaine, but just because LL isn't doing it, doesn't mean that AWG can't propose and implement the standard
[10:34] Susan Tsuki: it seems that a quarterly turn-around is pretty pessimistic. IETF internet drafts expire after six months. So if you submit something after a quarterly review, and they don't look at it for another 2.5 months, and they don't tell you how brilliant it is for another 3 after that, it can be pretty depressing. But I'm sure Zero only wanted to set a worst-case scenario; nothing prevents you for revising it with a new date and submitting it as a -02.txt in month 5. And emailing sldev again
[10:34] Zha Ewry: The IETF has a really good track record of what I'd call "Controlled balkanziatoin"
[10:34] Morgaine Dinova: Zha: if LL chairs the committee, then it WILL fly at the IETF. There's nobody to overrule the chair, if he wants to be despotic.
[10:36] Susan Tsuki: there's not a commitee, either, really, with a chair. A working group doesn't decide what gets promoted until after a bake off, which is very much not like a committee
[10:36] Zha Ewry: Chairs set some of the agenda, but they work in a pretty well defined process
[10:36] Susan Tsuki: a bake off is more like a LAN party
[10:36] Morgaine Dinova: Susan: issuing an OGP update every 3 months but only allowing 3 contacts with Zero in between updates pretty much ensures that nothing except LL's spec goes through.
[10:37] Latha Serevi: Another list of potential layers/RFC's summarized from Zha's messages: LLDL; Caps; Authentication; AD; Region Domain; Content and extensible low level prims; Asset serving; inventory serving. (hmm, where does physics fit in? scripting? object comms?)
[10:37] FWord Utorid: zha, when do you expect your happy ADto hit teh real world?
[10:41] Susan Tsuki: I would just say that of all the things M could have decided about open source, putting specs through the IETF probably has the best likelyhood of getting more interoperable code and extensions written. It puts LL in the position of Cisco to hacker's startups -- the real threat is creative talent getting bought up and I can think of worse things
[10:43] Morgaine Dinova: Zha: you've built that problem yourself, by making it your own project instead of an open one. It's a bad way to go. How come the AD is not an open project?
[10:44] Zha Ewry: It will be on gridforge as soon as I get the last bits cleared by my interneal processes
[10:44] Teravus Ousley: I should probably get going .. See ya at the zero meeting. :)
[10:44] Latha Serevi: Zero's office hours might be political chaos pretty much every month; shall we use these groupies hours as a way for us in the FIC to supplement those? Are there any other secret meetings that might be a better for us to scheme our secret agendas?
[10:45] Saijanai Kuhn: there's the opensim meeting at WRight Plaza
[10:45] Zha Ewry: I am not going to cancel thse as long a sthey are productive
[10:45] FWord Utorid: secret meetings don't work, use weapons and money and sex as your mechanisms for subterfuge
[10:47] Zha Ewry: "The likely to be called MMOX IETF based WG's maililng list"
[10:48] Morgaine Dinova: You're kidding .... LL are positioning this work under the "MMO" banner?
[10:48] Latha Serevi: You don't think we're very "M"? True enough.
[10:49] Zha Ewry: I think that is the ambition still
[10:49] Hermit Barber: EUTOVW might be more descriptive.
[10:49] Zha Ewry: Granted 83,000 concurrent falls a wee bit short
[10:49] Hermit Barber: Every user their own virtual world :-)
[10:50] Morgaine Dinova: It's just that LL have always shied away from anything that hints at "game", despite Philip's early statement that SL would subsume game worlds. So kind of amazed they'd be repositioning.
[10:50] Zha Ewry: Well, its the G that people get ithcy about
[10:51] Susan Tsuki: if you want to start a working group, it really helps to have a bunch of drafts submitted first, but the instructions are at [4]
[10:52] Zha Ewry: Susan, you can assume a request has been submitted
[10:54] Morgaine Dinova: I wonder what Eddy's position is going to be. Intel doesn't really have a vested interest in the space so he might be mainly representing Opensim. On the other hand, they may not even pay him to attend.
[10:54] Zha Ewry: Well, note that IBM views OpenSim as somethign we care about a lot
[10:55] Zha Ewry: (and the community, not just the slice of it we push on)
[10:55] Zha Ewry: I fully expect that OpenSim, in various forms is a huge (but not sufficient) part of the puzzle
[10:56] Susan Tsuki: bakeoffs almost always seem to have at least one remote participants in some working groups at least, because for example for networking protocols you want to have someone with a loopback reflector several hops aways
[10:57] Morgaine Dinova: Yeah, but it's a matter of focus. Opensim would have benefitted from early speccing of a "unrestricted interop" bit in objects, but it's not something you supported because you don't have a business case for it.
[10:57] Morgaine Dinova: So "caring about" is not always good enough.
[10:57] Zha Ewry: Umm. Code it Morgaine. Nobody is stopping you
[10:58] Hermit Barber: Morgaione: But a much beter answer is a structure on the wire that transfers an implementable licence set for anything. How that is dealt with at the server and the client then is an implementation detail.
[10:59] Morgaine Dinova: Gareth's work was totally stomped on by the lack of an interop bit so that he could freely transport those objects so marked.
[10:59] Morgaine Dinova: Hermit: please don't suggest that we put lawyers into the on-the-wire spec.
[11:01] Morgaine Dinova: There's no better "license" that says "This is unencumbered, if the author asserted it".
[11:01] Zha Ewry: I think we all agree, we want that, among others
[11:01] Zha Ewry: That is totally about Linden Lab business policy
[11:02] Hermit Barber: Morgaine: I can think of at least 3 seperate issues that might crop up. Look up creative commons if yoiu are unfamilar with this. But please don't jump on my head, I didn't deserve that and on a better day you would know it.
[11:02] Susan Tsuki: the essential part of a bakeoff is the scorecard, where participants decide on advance that there will the 10 points for each interoperable implementation of some list of a handful of specific major features, 25 points for all the essential things, and a few 5 pointers for the things that people aren't so confident about
[11:03] Susan Tsuki: I just made those numbers up, [5] has the original bakeoff scorecard example
[11:04] Latha Serevi: We all wish that LL would put in an "unencumbered" bit or some way for creators to mark their objects as free; but assuming they don't, can anyone suggest an expedient way to hack it in?
[11:04] Hermit Barber: Latha: Have a "Give me the license" plug-in.
[11:04] Morgaine Dinova: Hermit: We're all a bit jumpy because of LL's latest moves. But if you'd had been reading transcripts, you'd know that I've been pushing LL to adopt a single-bit CC license or a 3-bit one (one for each current CC license type) for the last 4 months.
[11:05] Susan Tsuki: if they submit something to the IETF, they are obligated to disclose encumberances as soon as they learn about them, and the IETF makes sure that the working group mailing list knows about them
[11:05] Hermit Barber: Then having got the license from the repo use it. A null response can be interpreted to be the most restrictive that still permits use.
[11:05] Hermit Barber: ie the same as LL asserts over user content in their TOS
[11:07] Latha Serevi: Erm, off the top of my head cheesy hack for in-world SL objects - a magic string in the description field meaning "unencumbered"? Decent idea, or terrible?
[11:07] Susan Tsuki: you can't apply for a patent and then file an internet draft that doesn't mention the patent. If you file the patent frist, likely people are going to ask about licensing before they set out implementing, and if the answer is too expensive or not forthcoming, it's usually a waste for them to even have submitted the draft because nobody will implement it. If it's a "reasonable" license, maybe some small percentage of the gross of any product using the claimed idea, then everybody has to be offered the same terms ("non-discriminatory")
[11:07] Hermit Barber: No Latha. I can change the description field of items I didn't make.
[11:08] Morgaine Dinova: Latha: as long as it allows in-world creators to enable it easily, and as long as interop gateways don't have to go through hoops trying to unravel LL's currently impossible-to-understand perms, then I'm all for it.
[11:08] Zha Ewry: its, 99% a "inside Linden's shop" issue, how to decie that
[11:08] Saijanai Kuhn: Lath, just a non-mod notecard
[11:08] Hermit Barber: But one person's "reasonable" is another persons laughable.
[11:08] Saijanai Kuhn: not sure if you can read non-mod
[11:09] Hermit Barber: But if there is a CPAN equivalent, put the license there.
[11:10] Hermit Barber: Now if creators want their products protected on other virtual worlds, they will add them to the CPAN repository.
[11:10] Morgaine Dinova: I'm actually very worried about the interaction between LL perms and object handling in OGP. Even LL acknowledges that SL perm semantics defy space and time ... expecting object gateways to work off their perms set is I think just totally unworkable.
[11:10] Hermit Barber: In that way you bypass LL entirely and they need to worry about how become compatible with yours.
[11:11] Morgaine Dinova: OGP should perhaps define its own, extensible, perms vector --- like the Unix perms but extendable.
[11:11] Susan Tsuki: I don't think it's very likely that you'll get anything more than a few patent applications out of LL. Currently they have 0 applications pending
[11:12] Latha Serevi: Thanks, Hermit, for the license-server approach, and Sai, for the no-mod-notecard fix to my hack.
[11:13] Infinity Linden: @Morgaine... yes... in the long run it would be nice to have a permissions declaration language where different virtual worlds could describe the semantics of their permissions... but... that is way far out in the future
[11:13] Saijanai Kuhn: the current copyright language proposal is kinda cumbersome
[11:14] Saijanai Kuhn: think its an MPEG thing will track it down
[11:14] Infinity Linden: MPEG/21 is designed for DRM. we don't do DRM.
[11:14] Morgaine Dinova: Well I think there are only two approaches to extensibility: either make the protocol itself extensible, or make the protocol support plugins that deliver sub-protocols. I don't mind which, as long as we're not hardwired to SL's set.
[11:15] Morgaine Dinova: Hi Infinity. No, sorry, "in the long run" we are all dead. Not going to put it off.
[11:15] Zha Ewry: At the protocol level, there's only one, really, which is the protocol is extensible.
[11:16] Infinity Linden: I look forward to reading your ID then
[11:16] Latha Serevi: Sai, shall we encourage someone to define a machine-readable fingerprint and standard notecard name for the unencumbered hack, so there can exist ASAP one way for an object creator to specify a freely distributable object?
[11:17] Saijanai Kuhn: a subset of it might be useable in this context, Infinity
[11:19] Zha Ewry: Just dazwed evryone into silence
[11:19] Infinity Linden: mmm... i might recommend DSI or SecPAL as a starting point
[11:20] Saijanai Kuhn: must be the other client making the typing sounds then
[11:20] Susan Tsuki: Infinity, I see your OGP trust model includes a section entitled, "the permission system should be suitably expressive"
[11:20] Infinity Linden: thought DSI has been described as "unfortunately academic" and SecPAL is a Microsoft product, both are more expressive over the full range of objects, transitivity options and permissions we would be interested in
[11:21] Morgaine Dinova: The vast bulk of items on the Internet are effectively "Copyright irrelevant", because despite all having copyright, only a few insane people try to deny access to things they make public. In the eventual metaverse, it has to be that way too, otherwise there won't be web-style free passage between worlds, which makes the current obsession with object copyright largely a waste of time.
[11:22] Saijanai Kuhn: I think the issue is of resale and so on, as far as LL content creators are concerned
[11:23] Infinity Linden: right... there is the inevitable problem that without DRM, it's hard to present certain types of IP (textures, scripts on web pages, etc.) without making it easy for people to illicitly copy
[11:23] Zha Ewry: /meI have a 11:30 as well, so off I go
[11:24] Latha Serevi: Anybody want to "take copy" of my blue sphere and see if its license file is vaguely along the right lines? (no-mod notecard, person wanting to take off-grid must check if creator is same as object creator?) Pls advise if I'm being an idiot.
[11:25] Morgaine Dinova: Infinity: indeed, as the music industry has learned. And instead of trying to buck it, find a new business model that accepts that reality.,
[11:25] Infinity Linden: so while i don't think we'll ever "solve the problem" of the risk of illicit copying of copyrighted materials... there are things you can do to make it less easy for the copier and certainly persisting license information in digital assets is a "must do"
[11:25] Hermit Barber: Have plugins that receive "Permission Requests." All the plugins receive the request and respond associatively with the permissions they can grant for that object identified by a UUID. No response means that permission isn't granted. The queriant asserts a cryptographic id and the responder does the same. The plugin is also responsible for providing a textual version of the permission on request.
[11:25] Hermit Barber: Now you have something that can get as many plugins as there are models.
[11:25] Latha Serevi: Mostly I want to know if this is forgeable.
[11:26] Morgaine Dinova: Well devolving all perms handling into a separate subsystem is certainly a good idea, otherwise OGP will be hardwired to ancient history.
[11:27] Saijanai Kuhn: I'm going afk for a while. Still recording transcript. Give me a clue when it's no longer Gorupies Time
[11:27] Morgaine Dinova: That doesn't necessarily have to be through plugins though, especially with LL getting agitated every time they're mentioned.
[11:27] Infinity Linden: hmm... now that Verisign is buying Certicom... we might get hybrid RW/ECDSA certs. that might even allow us to have digital signatures no a bunch of things we might not haev thought about in the past
[11:28] Hermit Barber: Now have a permissions UUID for any protected object.
[11:30] Infinity Linden: that you have so many plugins, that you can't do anything in the virtual world without having to download yet another plugin and potentially restarting
[11:30] Infinity Linden: plugins for supporting different auth techniques between a viewer and an agent domain.. that's probably a good idea.
[11:31] Morgaine Dinova: Well Zero brings up "balkanization" FUD every time that plugins are mentioned, so effectively yes, you're jumping up with evilness FUD at every opportunity. They're not evil, they're a mechanism for avoiding decisions being hardwired in.
[11:31] FWord Utorid: unfortunately, I have to restart too frequently already due to viewer crashes, sims restarting, or other built in problems
[11:31] Infinity Linden: @Fword.. surely you jest... there have never been problems with the virtual world
[11:32] Infinity Linden: hmm... forgot to add the <sarcasm> tags around that last one
[11:32] FWord Utorid: sighs... not as long as you're here :D :P :O
[11:32] Hermit Barber: Infinity: When the cache is properly designed then it really becomes non-problematic. I can understand that concern being real while it still wasn't possible to even cache graphics and textures properly, but surely by now it is "well understood to those practiced in the art" :-P
[11:34] FWord Utorid: the parts that don't work. like, the clothing. people in virtual worlds should not be allowed to wear clothing. this *is* a design flaw
[11:34] Latha Serevi: No URL yet, Morgaine, just a hint that it might be worth defining one.
[11:34] Latha Serevi: Haven't heard any "it's broken" comments yet, but nor have I heard any comments to the effect of "good plan, let's define such a notecard as an opt-in way of making your objects freely distributable". Yea? Nay?
[11:34] Infinity Linden: no... seriously... how would you improve the texture cache given it's existing limitations
[11:35] Infinity Linden: (though i'm now thinking that being able to allocate all remaining disk space to the cache would be good)
[11:35] FWord Utorid: well, the texture cache is basically optimized. using jp2k and allowing the staged downloads is all good
[11:35] Morgaine Dinova: Latha: I'm all for anthing that gives early interoperability, and your idea sounds feasible. Currently LL won't give us interop until they first solve world hunger and poverty. It's a sham.
[11:35] Infinity Linden: well... FWord... we _are_ in a mature sim... there's nothing keeping you from implementing your "no clothing" option
[11:36] FWord Utorid: goes back to the Calculatron 12k and asks once again what the best way is to divest people of their clothes
[11:37] FWord Utorid: maybe you need to try my new... ahem... invisible suit, yeah that's it!
[11:37] Morgaine Dinova: You mean your clients actually display people's clothes? How retro.
[11:38] Infinity Linden: no thanks... but i'll certainly give you extra points if you can get anyone in the "Emperor Class" to wear it... (where Emperor means any CEO or recognized Political Leader)
[11:38] FWord Utorid: yeah, points... at the expense of my head
[11:38] Morgaine Dinova: It's all in the eye of the beholder .... literally in this case.
[11:39] Hermit Barber: Ah, I would adopt a fractal compression stream and store and transmit items as small packets of base image plus multiple layers of corrections. I would reclaim top layers (higher frequencies) of corrections first. I would make the database distributed with simple associative caches so that anyone not having a texture could ask the sim, and the sim which should know the bandwidths available to each client could then ask clients to distribute the data to one another so unloading the server and providing greater bandwidth as more clients connect.
[11:39] FWord Utorid: hermit, the first sentence is how it works
[11:40] FWord Utorid: though i suspect it's more lzw than fractal
[11:40] Infinity Linden: cool idea Hermit.. we don't use fractal compression.. but otherwise do something similar
[11:40] Infinity Linden: i wonder if there's work here to instrument the cache
[11:41] Infinity Linden: the problems we've had in the past are that there really are a metric butt-load of textures out there
[11:41] Hermit Barber: Most with far too much resolution.
[11:41] FWord Utorid: one thing i would recommend is finding duplicates and weeding them and repointing uuids
[11:41] Latha Serevi: "I would reclaim top layers (higher frequencies) of corrections first" --- wouldn't the top layers be the lower frequencies, i.e. big chunky bits, not fine detail?
[11:41] Hermit Barber: And not distributed between clients. An hierarchical model. Not P2P at all.
[11:42] FWord Utorid: i am sure there are thousands of copies of the same texture in the asset db
[11:42] Infinity Linden: i believe we have three levels of resolution at the moment... gross, not so gross, and friggin detailed
[11:42] Morgaine Dinova: Hermit: it's only the last part of that (the P2P distribuition) that isn't done ... and for that you need unencumbered objects. Which is why we've been asking for the "unencumbered bit".
[11:42] Infinity Linden: i wonder if there's a way to try to tune the ratios of "gross" to "friggin detailed" you try to maintain
[11:43] FWord Utorid: actually, i wonder if the three levels makes the sim slower and if there would be the potential to turn that off
[11:43] Infinity Linden: i mean.. i wonder if you could walk around in five or six sims and tell the viewer... "okay... those are the sims i go to a lot... see if you can tune the cache parameters so things load faster"
[11:43] FWord Utorid: in theory you are downloading the equivalent of three textures per texture
[11:44] Morgaine Dinova: You're talking about this LOD tuning, when the last part of what Hermit said would provide massively more gain.
[11:44] FWord Utorid: i'm talking about killing LOD, for fun
[11:44] Infinity Linden: the "gross" texture is actually 1/9th or 1/16th teh size of teh "friggin detailed" image
[11:44] Latha Serevi: Infinity, I'd rather automatically collect some rough "where I spend time" or "where I teleport to" stats, in my client, and have that inform my local cache.
[11:45] Infinity Linden: i am talking about things i'm not sure about though... so don't quote me
[11:45] FWord Utorid: infinity, it's ok. the virtual world is all about bullshit
[11:45] Infinity Linden: hmm.. latha... that's a cool idea... lemme bounce it around and see what the peeps responsible for the rendering pipeline say
[11:46] Morgaine Dinova: Well it's no business of the server's to prejudge what the client wants sent.
[11:46] Morgaine Dinova: For scalability, the server should send NOTHING unless requested.
[11:46] Infinity Linden: @FWord... tsk tsk... about the only reply i can think of is ... "well... you're soaking in it!"
[11:47] Hermit Barber: Infinity, according to my husband there was a company called Iterated Systems that did a vast amount of theoretical and practical work on that.
[11:47] FWord Utorid: kinda sorta. maybe it's more like palmolive... softens hands while you soak in bullshit
[11:49] FWord Utorid: for people who like soaking in it
[11:50] Morgaine Dinova: Talking about limbo, where is the "Your account is not available for 5 mins" implemented, server or client? Because it kicks in when there is no reason (eg. broken client) for it, and needs to be stomped.
[11:50] Infinity Linden: no.. cornfields are for bad people.. maybe when we have system failures we could god teleport people to a cluster of OpenSim regions running on the four core under my desk
[11:50] FWord Utorid: infinity, bad example, term cornfield has a bad semantic. erase and substitute with 'limbo'
[11:51] Infinity Linden: yeah... maybe just call it "Inferno" and be done with it
[11:51] FWord Utorid: but if you teleported us into a computer under your desk, we might be looking... up
[11:52] Infinity Linden: Oh... i'm never at my desk.. just ask my boss
[11:52] FWord Utorid: ok. then the idea sucks. we need to be teleported somewhere that follows you around, so there could be a horde of people connected to your backpack
[11:53] Morgaine Dinova: I guess we need Zha's AD then to overcome LL's 5-minute block. This assumes that their support for interop goes beyond only allowing their own AD.
[11:53] FWord Utorid: morgaine: the 5 minute block isn't real
[11:53] FWord Utorid: well, not really. what happens is your account gets stuck somewhere
[11:53] FWord Utorid: if you switch login servers you can sometimes route around it
[11:54] FWord Utorid: my bots do that and rarely hit the 5 minute
[11:54] FWord Utorid: basically, your account is in a garbage collection cycle of some sort
[11:54] Morgaine Dinova: I rarely hit it too, but I missed 5 mins of this talk just because LL felt like kicking in the block for no reason.
[11:55] FWord Utorid: morgaine: like I said, it's account garbage collection. an improper disconnect seems to cause it frequently
[11:56] FWord Utorid: as for the 5 minutes you might have missed, sai does post logs religiously
[11:56] Infinity Linden: yup. improper disconnect is the most likely cause. and while i would hate to have people think I think SIP is the be-all, end-all presence solution... putting something more SIP like in OGP would probably be "a good thing"
[11:57] Morgaine Dinova: Yeah, it's odd. A plain crash doesn't usually make it kick in, but sometimes a rapid app closure does.