User:Whump Linden/Office Hours/Transcript 20080820

From Second Life Wiki
Jump to: navigation, search

Agenda

  1. Protocol Versioning
  2. Region Discovery
  3. Coordination for upcoming viewer and protocol changes

OGP Versioning Roadmap

  • Short Term
    • Versioning will be Out of Band
    • Version for all resources (Agent Domain, Region Domain, Region) will move together.
    • Will require coordination.
  • Longer Term
    • A known URL to get the current version of the protocol.
    • No per-message/per-resource versions.

Region Discovery

How does a user find a new domain or region to teleport to?

  1. They could type in a region_url, and the client can get the region domain, which'll retrieve the map for the region domain
  2. They could type in a region domain url, which'll retrieve a map that can find regions by whatever key that region domain uses to find regions easily without having a url.

Need User Stories

Upcoming Viewer and Protocol Changes

  • New viewer to fix OGP issues, and changes to protocol.
  • Changes to AD and Login
  • This will require maintenance on the patch
  • Possibly running old and new Login, AD, and Regions during transition

Chat Log

  • [13:25] Teravus Ousley: well, I really want to figure out which of the representations is 'correct'
  • [13:26] sacha Magne: hmm, the good one ?
  • [13:26] Teravus Ousley: particularly with the 'path' at the moment.
  • [13:26] Saijanai Kuhn: Warhog where did you hear that?
  • [13:26] Teravus Ousley: the path that's documented.. is /resource/action
  • [13:26] Teravus Ousley: the path that's implemented... is /resource/resourceID/action
  • [13:27] Teravus Ousley: in example form that might look like the following..
  • [13:27] Warthog Jun: one moment
  • [13:27] Sendao Goodman: coughs quietly
  • [13:27] DrFran Babcock: [1]
  • [13:28] Ash Venkman: Hi.
  • [13:28] Saijanai Kuhn: hey
  • [13:29] Strawberry Fride: hi folks
  • [13:29] DrFran Babcock: /hello Sendao
  • [13:29] Sendao Goodman: hello hello
  • [13:29] Teravus Ousley: implemented: /agent/323d5625-273b-4e5f-86ed-7a9c5ba41702/rez_avatar/request
  • [13:29] Teravus Ousley: documented: /agent/rez_avatar/request
  • [13:29] Bartholomew Kleiber: hi, all
  • [13:29] Whump Linden: oops, pardon
  • [13:30] Teravus Ousley: .. so resolving the difference would be nifty :D
  • [13:30] Teravus Ousley: hey
  • [13:30] Teravus Ousley: how are you doing today?
  • [13:30] Saijanai Kuhn: [2]
  • [13:30] Saijanai Kuhn: agenda for meeting
  • [13:30] Whump Linden: Afternoon everyone.
  • [13:30] sacha Magne: hi
  • [13:30] Whump Linden: I see some new folks here, welcome.
  • [13:30] Sendao Goodman: thank you
  • [13:30] DrFran Babcock: waves to Mal
  • [13:31] Malburns Writer: Hi All
  • [13:31] Whump Linden: if you don't mind hitting the URL Saijanai's posted above.
  • [13:31] Malburns Writer: Hi DrFran!
  • [13:32] Whump Linden: that's got today's agenda and some intial notes
  • [13:32] Mirt Tenk: hi BlueWall
  • [13:32] BlueWall Slade: Hello
  • [13:32] Whump Linden: initial, even
  • [13:32] Graph Weymann: I'm A Capability Guy, invited by Saijanai Kuhn
  • [13:32] Whump Linden: Awesome, Graph, welcome.
  • [13:33] DrFran Babcock: Whump, love that avatar!!!!
  • [13:33] Whump Linden: Graph, did you want any time on the agenda today? I can put you on at the end?
  • [13:33] sacha Magne: i tought there was a smoking ban in the US ;)
  • [13:34] Graph Weymann: I'm not familiar with the technical specifics
  • [13:34] Graph Weymann: but say something and I'll tell you if it's Wrong :)
  • [13:34] Ann Otoole: no. can smoke. rational thought is banned though
  • [13:34] Whump Linden: lol, okay, getting started
  • [13:35] BlueWall Slade: "/345 bear <secret key>"
  • [13:35] Cane Janick: lol
  • [13:36] Whump Linden: also, Leyla Linden will be here at 2pm SLT to join the discussion on coordination of the protocol change
  • [13:36] Dale Innis: hello all various people!
  • [13:36] Whump Linden: but first, protocol versioning in general
  • [13:36] Malburns Writer: Hi Dale
  • [13:36] Cane Janick: Hey Dale
  • [13:36] Dale Innis:  :)
  • [13:37] Whump Linden: in the short term, until the end of the year, at least, we are going to be dealing with OG Protocol changes out of band, so we'll be announcing them in the various venues
  • [13:37] Whump Linden: when we rev, the revision will be for all the resources
  • [13:38] sacha Magne: Q: Did we allows some compatibility or each revision means all hte ressources (sfotware) shoud be change at once ?
  • [13:39] Teravus Ousley: finds the answer to that question on the web page presented in the beginning of the meeting.
  • [13:39] Whump Linden: we're aiming for newer viewers to be able to talk to older agent domains and regions, if possible
  • [13:40] Whump Linden: but yes, that means going from 1.0 to 1.1 on all resources
  • [13:40] sacha Magne: ok
  • [13:41] Graph Weymann: What's a resource?
  • [13:41] Whump Linden: we're trying to avoid an explosion of components and versions
  • [13:41] Whump Linden: Graph: loosley speaking, an agent and a region
  • [13:41] Teravus Ousley: while I'd like it to be something that could be rhobust.. It's reasonable at this point in the protocol.. to do oob versioning..
  • [13:42] Dale Innis: agrees
  • [13:42] Strawberry Fride: surely it's best, at this stage, to have an agreed upon set of point releases and synchronise?
  • [13:42] Saijanai Kuhn: resource is also the verb used to obtain a CAP in he protocol
  • [13:43] Saijanai Kuhn: https://wiki.secondlife.com/wiki/Open_Grid_Protocol#Resources.2C_HTTP_.26_REST
  • [13:43] Whump Linden: Strawberry, so are you looking for a roadmap of point releases for the next few months?
  • [13:44] Strawberry Fride: I'm just looking for minimal confusion
  • [13:45] Strawberry Fride: I'm not an expert on the protocol design here - I've been sitting in trying to get my head round it all - my thoughts just come from the head of a developer
  • [13:45] Whump Linden: so that's what we're doing with the announcemnts
  • [13:46] Whump Linden: yes, and we will be publishing the next version before the switch over so that people can update their agent and region domains
  • [13:46] Teravus Ousley: well, one thing that I note about the spec and the implmentation that's different.. is the path url
  • [13:46] Strawberry Fride: cool, thanks :)
  • [13:47] Whump Linden: Teravus: that is not a surprise at the moment, which part of the spec are you looking at?
  • [13:47] Teravus Ousley: /resource/resourceID/name is how it's implmemented currently..
  • [13:47] Saijanai Kuhn: its not really a path though. It just loks like one
  • [13:48] Teravus Ousley: the definitions of URL space, as opposed to the implementation
  • [13:48] Saijanai Kuhn: at best its a naming convention
  • [13:48] Teravus Ousley: for example.. rez_avatar/request
  • [13:49] Teravus Ousley: implemented: /agent/323d5625-273b-4e5f-86ed-7a9c5ba41702/rez_avatar/request
  • [13:50] Teravus Ousley: it's saying, request a cap for agent 323d5625-273b-4e5f-86ed-7a9c5ba41702 (my Aditi UUID)
  • [13:50] Teravus Ousley: .. now the simulator wouldn't really necessarily know what that UUID means.. it has no knowledge of any Aditi UUIDs
  • [13:51] Whump Linden: no, the sim doesn't know about the UUID, but in turn it's making a request to the AD once it gets rez_avatar
  • [13:53] Saijanai Kuhn: total local chat lag.
  • [13:53] Whump Linden: https://wiki.secondlife.com/wiki/OGP_Explained#Teleport
  • [13:53] Dale Innis: ( It does seem quiet! )
  • [13:53] Strawberry Fride: thought I was crashing there
  • [13:53] Whump Linden: no, that's me looking up references in the wiki
  • [13:54] Teravus Ousley: nods. Read that.
  • [13:54] Graph Weymann: I have to say, I'm suspicious whenever I hear talk about things like "the uuid of an avatar" -- overlarge authorities ...
  • [13:55] Teravus Ousley: I guess the idea behind it is.. we're working on an object in the avatar space. the objectID is the following.
  • [13:55] Whump Linden: okay, going forward, we'd like to get to the point where there's a well known URL for the AD and Region Domain where you can ask what verison of the protocol it's running
  • [13:55] Teravus Ousley: (well agent domain space)/agent/agentID/action
  • [13:56] Saijanai Kuhn: well it does seem odd to be sending a known UUID in the CAP
  • [13:57] Graph Weymann: so these IDs are names only within the agent domain's authority? (agent domain = proxy for the client etc. I assume)
  • [13:57] Teravus Ousley: well, to be clear.. that's the request, not the caop
  • [13:57] Teravus Ousley: 'cap'
  • [13:57] Graph Weymann: (stop me if this is too OT)
  • [13:57] Teravus Ousley: .. that's the request for the cap
  • [13:57] Whump Linden: Graph the agent_id is within the AD yes
  • [13:57] Saijanai Kuhn: AGent domain is authorizing agent and intermediary for avatar and the rest of the metaverse
  • [13:58] Graph Weymann: sure. the problem case would be if it's actually global, in which case a plain UUID/GUID gives you no way to say who really owns it
  • [13:58] Whump Linden: there's a whole other discussion about disambiguation across domans
  • [13:58] Teravus Ousley:  :D
  • [13:58] Graph Weymann: the way to disambiguate is to avoid having namespaces :)
  • [13:59] Teravus Ousley: guess this leads back to having a list of providers (ip spaces) that you accept requests from agent domains
  • [13:59] Graph Weymann: or rather, arrange so that the name is (somewhere back in a chain) defined by a public key...so the private key holder has the naming authority
  • [13:59] Graph Weymann: then you cannot have naming conflicts or namespace administration issues
  • [14:00] Teravus Ousley: maybe this is more a topic for the AWG/Zero's then the huddle :D
  • [14:00] Ann Otoole: well google lively already proved it is useless to go by name so a single one central point of failure for identity management will forever be required.
  • [14:00] Whump Linden: okay, any other questions about versioning?
  • [14:00] Ash Venkman: There's stuff you can do with names without a central identity manager
  • [14:01] Leyla Linden: hello all
  • [14:01] BlueWall Slade: Hi
  • [14:01] Malburns Writer: Hi Leyla
  • [14:01] Enus Linden: waves @ leyla
  • [14:01] Bartholomew Kleiber: Hi, Leyla
  • [14:01] Ash Venkman: but as Graph said, for some things, you do have to base them on cryptographic keys at some level
  • [14:01] Teravus Ousley: well, would it be adventagious if the versioning be implemented by the cap names returned in the cap request?
  • [14:02] Teravus Ousley: .. or maybe a portion of the LLSD body of the cap request.
  • [14:02] Ash Venkman: that'd be nice
  • [14:02] Whump Linden: cool, okay Leyla just got here, so I'm going to jump ahead to coordinating the upcoming changes for the viewer and protocol
  • [14:02] Teravus Ousley: nods and listens
  • [14:03] Xugu Madison: eep?
  • [14:03] Xugu Madison: ack, wrong window, sory!
  • [14:03] Leyla Linden: okay, so the main change is that the agent seed cap is now returned with a 200 status code
  • [14:04] Leyla Linden: instead of 302 with a location header
  • [14:04] Ann Otoole: is this a voice conference? seems rather disjointed for a text conference.
  • [14:04] Whump Linden: / Ann: it's text
  • [14:04] Dale Innis: This is what Office Hours are usually like. :)
  • [14:04] Sendao Goodman: I think there's some hidden subtext heh.
  • [14:04] Saijanai Kuhn: extreme lag it seems
  • [14:05] Leyla Linden: but so that means we have too coordinate our viewer release with our server update
  • [14:05] Saijanai Kuhn: There are two groups that might be used for subtext and they are both quiet: Gridnauts and AW Groupipes
  • [14:06] Sendao Goodman: ^)
  • [14:06] Lillie Yifu: fiddles around with her key chain but realizes she doesn't have a secret key.
  • [14:06] Enus Linden: leyla: that also means that other consumers of the protocol ought to have advanced opportunity to accomodate, or, mutiple environments available to work against (old protocol implementation adn new), right?
  • [14:06] Enus Linden: e.g. opensim instances
  • [14:06] Whump Linden: Yes.
  • [14:06] Enus Linden: pyogp
  • [14:06] Enus Linden: gareth's world of wonder
  • [14:07] Enus Linden: etc
  • [14:07] Teravus Ousley: Leyla, I have a request.. that you may also be doing.. but I wanted to reiterate it.. when the updated viewer source comes out, could you please also have a binary release? Rules to OpenSim contribution say that I cannot look at the client source.. even if it's to simply build an OGP client from source and, that's been a limiting factor in my implementation of OGP in opensimulator in the past.
  • [14:07] Leyla Linden: ideally we'd put up the new auth alongside the old, so they'd both be accessible for testing as we transition
  • [14:08] Leyla Linden: Teravus, definitely
  • [14:08] Leyla Linden: i'll put them up on s3 and send out an email
  • [14:08] Teravus Ousley: Thank you muchly
  • [14:08] Whump Linden: Teravus, there's a binary release of the Open Grid Beta viewer at https://wiki.secondlife.com/wiki/Open_Grid_Public_Beta/Open_Grid_Beta_Viewers
  • [14:09] Teravus Ousley: fav's it.
  • [14:09] Infinity Linden: hope i didn't land on somebody
  • [14:09] Sendao Goodman: smiles
  • [14:09] BlueWall Slade: on the opensim side - the binary should include the file svn_revison with the release version or leave the .svn directory in the bin directory
  • [14:10] Signe Magic: hey yall
  • [14:10] BlueWall Slade: then we can get the rev from teh SL client if we need to
  • [14:10] Signe Magic: whats up?
  • [14:11] BlueWall Slade: our hopes
  • [14:11] Dale Innis: ( Is Whump's Office Hours! Talking about Things! )
  • [14:11] Signe Magic: mine are dead
  • [14:11] Saijanai Kuhn: discussing design of second life
  • [14:11] Signe Magic: oh cool
  • [14:11] Teravus Ousley: will be making changes in OpenSimulator to make it work well with OGP with less work getting it running.
  • [14:11] Lillie Yifu: And why the number 42 needs to be hard coded as a constant in the compiler...
  • [14:12] Infinity Linden: (actually it's the number 41 that's hardcoded)
  • [14:12] Saijanai Kuhn: 42 is the counting of the ImrpovedInstantMessenger options...
  • [14:12] Cane Janick: oh not the 0 thing
  • [14:12] Signe Magic: so you are all into the whole programming thing and that? myself i am a few months noob at flash ha ha
  • [14:12] Teravus Ousley: notes that 42 is also the answer to the meaning of life and everything.. but that's a bit outside the scope of this discussion.
  • [14:12] Whump Linden: yes
  • [14:12] BlueWall Slade: a flasher eh?
  • [14:12] Dale Innis: Next agenda item.... :)
  • [14:13] Signe Magic: i try to learn software as much as possible
  • [14:13] Ann Otoole: is 42 someone's age? will next year it require 43?
  • [14:13] Whump Linden: so would it be helpful to the OpenSim side if we ran the old and new versions in paraell for a a period of time
  • [14:13] Signe Magic: but you know i dont spend all my time with software
  • [14:13] Saijanai Kuhn: 42 is the answer to the question posed by the extra-dimensional white mice, Ann
  • [14:13] Enus Linden: 42 is the answer
  • [14:14] Sendao Goodman: o.o
  • [14:14] sacha Magne: Thanks for the fishes
  • [14:14] Enus Linden: to the meaning of [lifehttp://en.wikipedia.org/wiki/The_Answer_to_Life,_the_Universe,_and_Everything]
  • [14:14] Ann Otoole: python likes white mice for lunch
  • [14:14] Enus Linden: i wont bother fixing that link
  • [14:14] Signe Magic: i make those silly little dress-up games and mini animations ha ha... i would like to learn 3d modeling better, but havent found a good program
  • [14:14] Harleen Gretzky: lol
  • [14:15] Whump Linden: / Signe, at the moment, we're talking about a protocol for virtual world interoperablity
  • [14:15] Signe Magic: bear?
  • [14:15] Signe Magic: cool
  • [14:15] Signe Magic: sorry if im disturbing heh... i dont know these things, i wish i could learn though
  • [14:16] Dale Innis: Lots of opportunities to learn! :)
  • [14:16] Teravus Ousley: Signe: it's a long read but: https://wiki.secondlife.com/wiki/Open_Grid_Protocol#Resources.2C_HTTP_.26_REST
  • [14:16] Lillie Yifu: Then sit and listen. I had a question. Right now we have agent domains handing off between regions. Tao's AD implementation asks the question of whether there are plans to allow ADs to ahnd off to other ADs.
  • [14:16] Infinity Linden: no
  • [14:17] Infinity Linden: there are no such plans
  • [14:17] Whump Linden: okay, so going back to the earlier question, should we leave the old version up along side the new, or roll the new version and let folks catch up?
  • [14:17] Infinity Linden: if you have an account / agent / inventory that lives in one agent domain
  • [14:17] Infinity Linden: it stays there
  • [14:17] Infinity Linden: but
  • [14:17] Dale Innis: Having both up seems useful; is there a downside to it?
  • [14:17] Infinity Linden: d'oh
  • [14:17] Enus Linden: leave the old version up imo whump
  • [14:17] Saijanai Kuhn: its a dangerous practice, IMHO, Lillie
  • [14:17] Infinity Linden: i'm getting ahead of myslf
  • [14:17] Enus Linden: tho
  • [14:17] Enus Linden: maintaing 2 envs is overhead
  • [14:18] Signe Magic: yeah well ill take it slow and learn these things more and more hopefully i will be able to learn to make a bit more complicated flash games and then... well it would take a while for me to learn anything really complicated ha ha
  • [14:18] Saijanai Kuhn: An AD is potentially the ultimate man in the middle attack on the rest of the system. The more AD's an avie is associated with the more dangerous and messy the situation (IMHO)
  • [14:18] Enus Linden: but, releasing docs and updated code to the beta deploy without advance opportunity to accomodate would halt testing for some period of time.
  • [14:18] Whump Linden: yes
  • [14:18] Teravus Ousley: thinks that there should be a time period that the old version is up.. and it shouldn't be more then a week or two after the announcement of the new version.
  • [14:18] Leyla Linden: the downside is that they would talk to two different auth services, and at some point there needs to be a switch to make the "new" the standard
  • [14:18] Infinity Linden: well... yeah... it is... but it allows us to respond more quickly
  • [14:19] Enus Linden: so we are currently only talking about auth changes
  • [14:19] Infinity Linden: the idea is that you have revision N of the spec and revision N+1
  • [14:19] Enus Linden: do we foresee region domain changes coming down the road too?
  • [14:19] Enus Linden: in the near future that is
  • [14:19] Dale Innis: waves while dashing rudely off; will read chatlog! :)
  • [14:19] Teravus Ousley: heh, rez_avatar/rez, vs rez_avatar
  • [14:20] Malburns Writer: i would do quick switch at this early stage - maybe week or two of overlap
  • [14:20] Infinity Linden: and you would have two distinct environments that are supposed to be running each
  • [14:20] Teravus Ousley: does his best to keep up :D
  • [14:20] Whump Linden: and there's a change to rez_avatar's message, the hostname instead of ip address
  • [14:20] Signe Magic: i will be afk for a few minutes... be right back, will try to humbly learn from your conversation :)
  • [14:20] Infinity Linden: so it might be that we have something like rev X of the Open Grid viewer and it works with environment A
  • [14:20] Enus Linden: which rez_avatar call whump?
  • [14:20] Infinity Linden: and then when we release a new rev of the spec
  • [14:20] Enus Linden: runs to docs
  • [14:21] Whump Linden: rez_avatar/rez
  • [14:21] Infinity Linden: we (eventually) releaes a new rev of the viewer with the underatndingi it's intended to be used in environment B
  • [14:21] Saijanai Kuhn: Signe Magic, this is a conversation that has been going on for a year
  • [14:21] sacha Magne: the Release_Channel ?
  • [14:21] Infinity Linden: the concerns are that a) it takes a little time between the releasing a spec and deploying an implementation of it
  • [14:21] Whump Linden: Enus: [3] and the related PJIRA
  • [14:22] Enus Linden: ty whump
  • [14:22] Saijanai Kuhn: Signe Magic: https://wiki.secondlife.com/wiki/Category:Grid_Interoperability_Chat_Logs
  • [14:22] Infinity Linden: and b) we have at least 3 different implementations of agent domain that have to be coordinated
  • [14:22] Whump Linden: Infinity, at the moment we only have a couple of implementations that need to rev
  • [14:22] Lillie Yifu: Could I be evil and suggest doing just a sharp cut over?
  • [14:22] Sendao Goodman: no.
  • [14:22] Lillie Yifu: I know people are all in favor of "keep old and new"
  • [14:22] Sendao Goodman:  :)
  • [14:22] Infinity Linden: from what to what?
  • [14:22] Lillie Yifu: but no one shuold be doing anything that is
  • [14:22] Lillie Yifu: hmmm
  • [14:22] Strawberry Fride: I say go with the new
  • [14:22] Lillie Yifu: depending n this
  • [14:23] Lillie Yifu: if we lose the capability for a few days and have a cleaner transition, that would be better
  • [14:23] Sendao Goodman: if you start with a good versioning cutover capability now you can continue with it later, yes? and it will be more important later
  • [14:23] Lillie Yifu: than a messy one.
  • [14:23] Infinity Linden: as it stands, there are new messages in draft 3 of the spec that are different from draft 1
  • [14:23] Infinity Linden: the problem with tht Lille is that the way a lot of people develop...
  • [14:23] Whump Linden: well the fallout will be that there are non-developers participating in the beta, so that means that they will be confused by the breakages and I get to explain to them why they can't get their bear at the moment. :)
  • [14:24] Saijanai Kuhn: seems to me that once pyogp gets workign as intended, pyogp should get the draft 3 first for QA purposes before its released for implementation by everyone else
  • [14:24] Teravus Ousley: nods. I'll be interested in that spec once it comes out :D
  • [14:24] Infinity Linden: they start developing featuers in a private branch of the code
  • [14:24] Infinity Linden: then integrate it iwth "the tip" or the release brahcn
  • [14:24] Infinity Linden: so if we did a sharp cutover, the people developing on private branches would need to integrate with the tip before they could expect to have thier enhancements testable
  • [14:25] Lillie Yifu: umm which would be good, so that we don't have the problem of the little grains of rice sneaking under the carpet.
  • [14:25] Infinity Linden: another approach is to have rev N of the code handle rev N and rev N-1 of the spec
  • [14:25] Infinity Linden: but that can get a touch dicey
  • [14:26] Whump Linden: though it gives us some resilience
  • [14:26] Infinity Linden: trying to remember which features are in which rev of the spec and which rev of the code
  • [14:26] Lillie Yifu: ummm I think that if infinity can't remember it
  • [14:26] Lillie Yifu: the rest of us are lost.
  • [14:26] Leyla Linden: hmm, doable, but more overhead
  • [14:26] DrFran Babcock: agrees
  • [14:26] Infinity Linden: no no... i mean when yer head down in the code
  • [14:27] Infinity Linden: you sometimes have to stop and say. 'oh heck... did we remove such and such a feature'
  • [14:27] Infinity Linden: and then you ahve to go look at the ref
  • [14:27] Infinity Linden: which is a problem you have when you have a lot of people developing in private branches and then merging into release branches
  • [14:28] Infinity Linden: so while it's possible to do, it tends to make you a less productive programmer (well.. i think that anyway)
  • [14:28] Signe Magic: back
  • [14:28] Leyla Linden: so a good question is, if we are keeping our spec updated with our code, how do we warn of new changes before it's deployed?
  • [14:28] Infinity Linden: right
  • [14:29] Infinity Linden: we do a round of internal review of the spec
  • [14:29] Infinity Linden: then start coding to it
  • [14:29] Saijanai Kuhn: and a round with pyogp
  • [14:29] Infinity Linden: then we release the spec
  • [14:29] Infinity Linden: and get other people to start coding to it
  • [14:29] Saijanai Kuhn: THE release the spec
  • [14:29] Teravus Ousley: expects at this point in the protocol definition that there will be times when it isn't feasable to have N & N-1, and other times it will be..
  • [14:29] Infinity Linden: and at some point, we have something we can make live
  • [14:29] Ann Otoole: who is to say LL is the authority for that matter. What if the open universe decides to leave sl behind?
  • [14:29] Enus Linden: the code should not be kept up to date with the spec imo leyla
  • [14:29] Whump Linden: we're running against the time limit, so is this something to continue in gridnauts?
  • [14:30] Infinity Linden: ann...
  • [14:30] Enus Linden: the spec should be iterated on, refined
  • [14:30] Infinity Linden: they;re more than welcome to walk away from our 1,000,000+ users
  • [14:30] Enus Linden: and then updates announced and published
  • [14:30] Saijanai Kuhn: the spec will be submitted to a standards body eventually
  • [14:30] Lillie Yifu: Suggestion, in the old spec have changes that are comign labelled
  • [14:30] Enus Linden: and then i may have another
  • [14:30] Enus Linden: oops
  • [14:30] Lillie Yifu: "deprecated"
  • [14:30] Infinity Linden: yeah.. Sai.. but for the time being... that's months (if not years) away
  • [14:31] Lillie Yifu: as soon as the decision is made, and before the new spec is released
  • [14:31] Lillie Yifu: that way people using the old spec will have warning before the spec is released.
  • [14:31] Signe Magic: what parts of second life is it you guys make?
  • [14:31] Saijanai Kuhn: right and no-one else but SL, OpenSim and realXtend are worried about this stuff in any menaning ful way, and they've all agreed to work together
  • [14:31] Infinity Linden: yup... Lille
  • [14:31] Infinity Linden: that's a good idea
  • [14:31] Infinity Linden: to date we've been coordinating with people on in ad hoc manner
  • [14:31] Ann Otoole: so invent a standards body and be done with it. everyone must submit to the standardss body. not too hard to make your own ISO segment
  • [14:32] Teravus Ousley: thinks that when a new version of the protocol is released, a week of overlap time would be great for implementing.. but does understand that it may not always be feasable.
  • [14:32] Graph Weymann: IETF!
  • [14:32] Infinity Linden: it will be a cold day in hell before i deal with teh ISO again
  • [14:32] Infinity Linden: or the CCITT
  • [14:32] Ash Venkman: yeah, the IETF seems like an almost-reasonable standards body in comparison :)
  • [14:32] Teravus Ousley: hehe
  • [14:32] Ann Otoole: lmao!
  • [14:32] Cane Janick:  :)
  • [14:33] Enus Linden: waves buh bye
  • [14:33] Whump Linden: yes, this sort of protocol is in line with things IETF vets
  • [14:33] Lillie Yifu: The lindens here, and some of the other people here are making parts of second life that don't exist yet, Signe. The rest of us are just here to find out what is going on and perhaps make useful suggestions about how to make the hanges as painless as possible. Or at least make it so if there's pain, it is fun pain.
  • [14:33] Enus Linden: ty whump
  • [14:33] Ann Otoole: well you know what i mean then. make up a standards body and be done with it.
  • [14:33] Ash Venkman: (it still hurts when I think about SIP though)
  • [14:33] Whump Linden: / thanks Enus
  • [14:33] Infinity Linden: i think we're definitely "philosophically aligned" with the IETF
  • [14:33] Infinity Linden: cheers E
  • [14:33] Teravus Ousley: Ty all
  • [14:33] Teravus Ousley:  :D
  • [14:33] Signe Magic: thanks for being patient with me. i'd love to learn scripts and stuff, i can barely understand actionscript.
  • [14:34] Infinity Linden: and if anyone has ever tried to implement X.500 or X.680, you know what pain emerges from design by committee
  • [14:34] Xugu Madison: Must run, take care all!
  • [14:34] Ash Venkman: Eek.
  • [14:34] Infinity Linden: ack
  • [14:34] Infinity Linden: i have to vanish as well
  • [14:34] Teravus Ousley: >.>
  • [14:34] Whump Linden: I'll have to follow up on the question about running two versions
  • [14:34] Lillie Yifu: A friend of mine says that a standard is a camel designed by a committee.
  • [14:34] sacha Magne: Gnight all
  • [14:34] BlueWall Slade: umm, what do they say? rought concensus and running code??
  • [14:34] Teravus Ousley: take care.
  • [14:34] Whump Linden: Any quick questions before we break off?
  • [14:34] Infinity Linden: okay... whump... we shold come up with a recommendation
  • [14:34] BlueWall Slade: see yah sacha
  • [14:35] Lillie Yifu: kk could we start having just a docs triage for this?
  • [14:35] Infinity Linden: but we have to make sure it works with opensim developers, OSGrid like operators and PyOGP
  • [14:35] Whump Linden: yes, we'e all monitoring #gridnauts so we can discuss there
  • [14:35] Infinity Linden: Lille
  • [14:35] Lillie Yifu: And yes I am volunterring to attend
  • [14:35] Infinity Linden: i'll take that as an action item to try to pin down a date for the release of draft 3 of the spec
  • [14:35] Saijanai Kuhn: #gridnauts is sorta the defacto OGP channel, I think
  • [14:36] Infinity Linden: and then to schedule a meeting (either here or at Zero's office hours) to discuss it
  • [14:36] Infinity Linden: yeah
  • [14:36] Infinity Linden: and it'll probably be announced on #gridnauts
  • [14:36] Lillie Yifu: kk
  • [14:36] Saijanai Kuhn: and spammed elsewhere ;-)
  • [14:36] BlueWall Slade: hehe
  • [14:36] Whump Linden: yes, thanks Saij
  • [14:36] Lillie Yifu: looks around for that irc to sl bridge she's wanted fo rso long.
  • [14:36] Infinity Linden: oh yeah... .all i have to do is IM Sai... he does a good job of cross posting
  • [14:36] Saijanai Kuhn: Thraxis has one written in libsl, LIllie
  • [14:37] Infinity Linden: hmm... sounds like a business opportunity...
  • [14:37] Ann Otoole: looks for unhacked iirc and finds none
  • [14:37] Infinity Linden: okay
  • [14:37] Infinity Linden: cheers all
  • [14:37] Ash Venkman: lillie, there isn't one yet? I figured that'd be the first thing people would write
  • [14:37] Infinity Linden: thanks whump!!!
  • [14:37] Lillie Yifu: there is one for the text interface, but I don't know if a graphical client has one.
  • [14:37] Whump Linden: thanks Leyla, Infinity, for coming by to answer questions