User:Whump Linden/Office Hours/Transcript 20080910

From Second Life Wiki
Jump to navigation Jump to search
The printable version is no longer supported and may have rendering errors. Please update your browser bookmarks and please use the default browser print function instead.

Agenda

Today we released Draft 3 of the Open Grid Protocol. Tess and Infinity dropped in to talk about the updated spec and answer questions.

Chat Log

  • [13:22] Whump Linden: Hey, Teravus.
  • [13:23] Teravus Ousley: Afternoon Bill.
  • [13:23] Teravus Ousley: was reading the draft and didn't want to read it.. through the meeting.. by forgetting the time.
  • [13:25] Whump Linden: Well, Tess and Infinity will be here soon and they're ready to answer questions.
  • [13:26] Teravus Ousley: looks fairly similar in the region domain space
  • [13:27] Whump Linden: yes, the major change is reviewing the avatar data before handing it off to the new region.
  • [13:30] Saijanai Kuhn: HGey Enus, er, Whump
  • [13:30] Whump Linden: Hey Saijanai
  • [13:30] Whump Linden: Zha
  • [13:30] Zha Ewry: Where?
  • [13:30] Zha Ewry: Is that slacker back from vacation?
  • [13:30] Xugu Madison: wow, people at a meeting! IT's been a while...
  • [13:31] Whump Linden: Publish a new draft and everyone comes out of the woodwork.
  • [13:31] Xugu Madison: There's an important lesson here
  • [13:31] Xugu Madison: I'm just not sure what it is
  • [13:31] Zha Ewry: Publish early and often
  • [13:31] Infinity Linden: hola
  • [13:31] Zha Ewry: Oh, and publish in Chicago, where dead people can publish.
  • [13:31] Whump Linden: Hey, Tess, Infinity!
  • [13:31] Teravus Ousley: Hi
  • [13:32] Zha Ewry: Its raining lindens
  • [13:32] Tess Linden: hey :)
  • [13:32] Infinity Linden: yay! the gang's all here!
  • [13:33] Whump Linden: Okay, so the new draft of OGP has gone up on the wiki: https://wiki.secondlife.com/wiki/Open_Grid_Protocol
  • [13:34] Whump Linden: At this time, we've not made changes to the Agent Domain, or our Regions.
  • [13:34] Infinity Linden: and of course you've all been able to review it, so we'll have three minutes of discussion followed by ratification (joke)
  • [13:34] Teravus Ousley: hehe, so we'll change it to rez_avatar/rez on the region domain..
  • [13:34] Xugu Madison: First thought; there's at least two docs on LLSD now, https://wiki.secondlife.com/wiki/OGP_LLSD and https://wiki.secondlife.com/wiki/LLSD
  • [13:34] Infinity Linden: and soon there will be one more
  • [13:34] Infinity Linden: WRT LLSD
  • [13:35] Infinity Linden: we're trying to unify all the docs
  • [13:35] Whump Linden: Yes, need a disambigugation page.
  • [13:35] Infinity Linden: into one doc
  • [13:35] Xugu Madison: Cool, as long as you know
  • [13:35] Infinity Linden: after which we say "everything else is depricated"
  • [13:35] Saijanai Kuhn: wow crazy visua artifacts right now. Non-stop flickering textures
  • [13:35] Whump Linden: Xugu: yes, thanks for pointing that out
  • [13:35] Infinity Linden: hmm.. is that 'cause we're talkinga bout L-LSD?
  • [13:36] Teravus Ousley: considers a JSON implementation..
  • [13:36] Infinity Linden: yeah.. .when you add the various descriptions of LLSD in the spec, in the wiki and the comments in the code, yeah... it was a touch confusing
  • [13:36] Infinity Linden: Teravus... that is coming
  • [13:36] Whump Linden: So, since we didn't have a lot of time for people to review before this meeting, maybe Tess or Infinity could do a quick run through of what's new?
  • [13:37] Infinity Linden: okay.. .want me to go first?
  • [13:37] Teravus Ousley: notes it doesn't appear to change 'much' from the region domain service perspective.
  • [13:37] Whump Linden: sure
  • [13:37] Infinity Linden: okay.. i think the base document did not change
  • [13:37] Infinity Linden: we introduced LLIDL
  • [13:37] Infinity Linden: or LL interface description language
  • [13:37] Whump Linden: Oh, and I should point out that we've split the document into four parts now.
  • [13:37] Infinity Linden: which is the formal specification for resource classes
  • [13:38] Infinity Linden: yup. very true
  • [13:38] Whump Linden: tries not to think of WS-*
  • [13:38] Teravus Ousley: >.>
  • [13:38] Teravus Ousley: <.<
  • [13:38] Infinity Linden: eek
  • [13:38] Infinity Linden: well... at least our documents are written such that the could also be combined as chapters in a single spec
  • [13:38] Infinity Linden: and who knows
  • [13:38] Infinity Linden: maybe that'll be done
  • [13:39] Infinity Linden: after internal review, we all kind of agreed that my first hack at the LLIDL section was weak
  • [13:39] Infinity Linden: not an epic fail, but i admit, it is not entirely clear
  • [13:39] Infinity Linden: so i'm in the process of rewriting it
  • [13:40] Infinity Linden: we're publishing this draft to "get something out there"
  • [13:40] Zha Ewry: chuckles, so a mildly heroic fail?
  • [13:40] Infinity Linden: nah.. i'm thinking really o the Rock Band ratings...
  • [13:40] Xugu Madison: LLIDL looks less like someone forced a standard into XML, at least :)
  • [13:40] Infinity Linden: like "Weak Solo" or something
  • [13:40] Tess Linden: but as you'll see in the teleport authentication sections, they now include llidl to give you more exact specs
  • [13:40] Zha Ewry: chuckles
  • [13:40] Infinity Linden: yup
  • [13:40] Zha Ewry: "You broke three strings, but you hand isn't bleeding"
  • [13:40] Infinity Linden: LLIDL is "the" standard for how we describe abstract data types
  • [13:41] Infinity Linden: XML is now just a serialization
  • [13:41] Infinity Linden: and after much discussion
  • [13:41] Infinity Linden: we're going to add "official" serializations for JSON and a binary encoding
  • [13:41] Xugu Madison: So, for example, we'll be seeing a RegAPI LLIDL then?
  • [13:41] Zha Ewry: Excellent
  • [13:41] Infinity Linden: it is entirely possible
  • [13:42] Infinity Linden: but i don't want to speak for that team
  • [13:42] Infinity Linden: it may be RegAPI or son of RegAPI
  • [13:42] Infinity Linden: but even if it's not, we could probably take the XML they use
  • [13:42] Infinity Linden: and back convert it to LLIDL
  • [13:42] Xugu Madison: They said we'd have news on son of, within a fortnight, so probably son of RegAPI
  • [13:43] Zha Ewry: chuckles, Bastard, unwanted child of RegApi
  • [13:43] Infinity Linden: yay! i got out of that without having to commit another team with an action item!
  • [13:43] Teravus Ousley: hehe
  • [13:43] Infinity Linden: but seriously... i'll chat with them about their plans
  • [13:43] Infinity Linden: but it's waaaaay premature to announce anything
  • [13:44] Infinity Linden: and RegAPI is not under my control
  • [13:44] Infinity Linden: so
  • [13:44] Infinity Linden: moving along
  • [13:44] Infinity Linden: since we may be short of time
  • [13:44] Teravus Ousley: agent info { // Under construction }
  • [13:44] Infinity Linden: we have a new auth scheme
  • [13:44] Infinity Linden: we've sort of formalized the account / agent relationship
  • [13:45] Infinity Linden: so you can authenticate yerself using an account credential
  • [13:45] Tess Linden: avatar info?
  • [13:45] Infinity Linden: if it's supported
  • [13:45] Infinity Linden: or an agent creential
  • [13:45] Teravus Ousley: in the vetted rez_avatar
  • [13:45] Teravus Ousley: avatar data: ? ; under construction
  • [13:46] Infinity Linden: mmm... the identifiers and credentials for auth are distinct from the avatar data
  • [13:46] Infinity Linden: and
  • [13:46] Infinity Linden: we added two new authentication schemes
  • [13:47] Infinity Linden: as there are some serious concerns about the use of MD5 in an environment where you ahve authenticators
  • [13:47] Infinity Linden: running between administrative domains
  • [13:47] Infinity Linden: in the past we got away with it 'cause all the MD5ed passwords went over SSL
  • [13:47] Teravus Ousley: nods
  • [13:47] Infinity Linden: but the message you should take away is
  • [13:47] Zha Ewry: Which gives you a fair bit of cover, SSL, being pretty solid
  • [13:48] Infinity Linden: it's entirely possible we will deploy new auth schemes in the future
  • [13:48] Infinity Linden: and they would probably look a lot like what you see in the spec
  • [13:48] Infinity Linden: that being said
  • [13:48] Infinity Linden: this is not an announcement of a new service
  • [13:48] Infinity Linden: os
  • [13:48] Infinity Linden: so
  • [13:48] Xugu Madison: In particular, if you're just sending the MD5 of the password, not MD5 of password + timestamp, aren't you vulnerable to repeat attacks?
  • [13:48] Infinity Linden: no comment
  • [13:48] Infinity Linden: and Xugu
  • [13:49] Zha Ewry: You are, Xugu, if its in plaintext
  • [13:49] Infinity Linden: security issues would take up ALL of the remaining time
  • [13:49] Infinity Linden: so
  • [13:49] Infinity Linden: maybe we could defer them for right now
  • [13:49] Xugu Madison: no problem!
  • [13:49] Xugu Madison: again, as long as you're aware
  • [13:49] Infinity Linden: one more thing before i handoff to tess
  • [13:49] Infinity Linden: login-time, per-user maintenance has been added to the spec
  • [13:50] Infinity Linden: so essentially what you might get from an auth request is (fail, success - here's a seed cap, or uhoh - here's a cap to a resource describing when i'm going to be finished iwth maintenance)
  • [13:50] Infinity Linden: and that's about it
  • [13:51] Zha Ewry: This is so you can do a arbitraty lenght maint step on login?
  • [13:51] Infinity Linden: quick questions before i hand off to Tess?
  • [13:51] Infinity Linden: technically yes
  • [13:51] Zha Ewry: (on a per account basi)
  • [13:51] Infinity Linden: and the idea is that the maint cap reference will have an estimate for how long maint will take
  • [13:51] Zha Ewry: We'll assume its normally miinutes or seconds, not days or weeks
  • [13:51] Zha Ewry: Does it loop on the maint cap?
  • [13:52] Infinity Linden: at the end of that period, you access the cap again and if it's really finished
  • [13:52] Infinity Linden: you get your seed cap
  • [13:52] Zha Ewry: Ahhhh
  • [13:52] Infinity Linden: (or another maint cap)
  • [13:52] Infinity Linden: but yeah
  • [13:52] Zha Ewry: So you can loop on it, and if it wants it could update the status
  • [13:52] Infinity Linden: the spec says maint can go on forever
  • [13:52] Infinity Linden: but i think if we did that
  • [13:52] Infinity Linden: we would lose a lot of residents
  • [13:52] Infinity Linden: os
  • [13:52] Infinity Linden: os
  • [13:52] Infinity Linden: so
  • [13:52] Infinity Linden: we're not going to do it
  • [13:53] Infinity Linden: and now i hand off to Tess...
  • [13:53] Tess Linden: the spec just allows the option for maintenance
  • [13:53] Tess Linden: so Teleport. Notice we hadded a Vetted Avatar Data section
  • [13:54] Zha Ewry: accepted your inventory offer.
  • [13:54] Tess Linden: the current protocol trusts sim A to serialize your avatar to sim B, but who's to say sim B would transform your avatar completely, and teleporting back to sim A would overwrite what you'd like it to be?
  • [13:54] Teravus Ousley: technically, that can come from a simulator also though.. in the case of derez
  • [13:55] Teravus Ousley: right.. or send null.. and .. ooops where'd you go?
  • [13:55] Tess Linden: The Teleport Sequence section now explains the UML diagram pretty much but in words
  • [13:56] Tess Linden: We included notes for everything thats deployed differently
  • [13:56] Tess Linden: couple of big key changes
  • [13:56] Tess Linden: 'region_url' is renamed to 'public_region_seed_capability'
  • [13:57] Tess Linden: and seed_capability is renamed to region_seed_capability to distinguish from agent_seed_capability
  • [13:57] Tess Linden: also 'rez_avatar/rez' instead of "rez_avatar"
  • [13:57] Teravus Ousley: :D
  • [13:57] Graph Weymann: personally I'd call it seed_cap...when everything is a capability, talking about it too verbosely makes it hard to think about :)
  • [13:58] Tess Linden: Graph: we do have to distinguish between the different seed caps because only certain cap names can be granted depending on who's handling the seed
  • [13:58] Whump Linden: And again, we've not deployed this yet. We're working out the plan for that.
  • [13:58] Graph Weymann: Sure, I mean don't spell out "capability" -- adds verbosity with little info
  • [13:58] Graph Weymann: for that matter, just call them foo_seed
  • [13:58] Graph Weymann: unless you have seed corn as well as seed caps
  • [13:59] Tess Linden: Notice the change in formatting of the entire document. For each suite, we describe the resource and provide the LLIDL for each specific interface
  • [13:59] Zha Ewry: it would be nice, if we could have an overlaping deployment, so that new path code can be tested before we break old path
  • [13:59] Teravus Ousley: It's much more clear.
  • [13:59] Infinity Linden: thinks of where we might sew corn in the spec
  • [13:59] Stu Paulino: So what does everyone do here
  • [14:00] Tess Linden: Zha: yes, thats the plan
  • [14:00] Mirt Tenk: we're currently in a meeting Stu, working on code
  • [14:01] Teravus Ousley: the changes are so similar as far as region domain goes.. it could be done simply as far as that's concerned.
  • [14:01] Teravus Ousley: (two versions)
  • [14:01] Zha Ewry: I think that mighrt be nicest, but if we roll it onto a few sims first, that would be tolerable
  • [14:01] Stu Paulino: no probs guys, over my head anyway. Just exploring
  • [14:02] Whump Linden: and we could put out a separate AD that one could call from a command line argument
  • [14:02] Teravus Ousley: wonders if he could also run them simulataniously for a while on the RD
  • [14:02] Stu Paulino: I willleave you to ot
  • [14:02] Teravus Ousley: will have to experiment mentally :D
  • [14:02] Whump Linden: / no worries, Stu
  • [14:03] Infinity Linden: i don't want to speak for the ops team or for the folk who will be doing the work... but that sounds like a plan
  • [14:03] Infinity Linden: we've had different code on login0 and login1 in the past
  • [14:04] Tess Linden: There's a big change that isn't obvious. The region_seed_capability that you get back from rez_avatar/request is also what is used to request derez with
  • [14:05] Tess Linden: region_seed_capability is essentialy tied to this particular agent for this particular session
  • [14:05] Zha Ewry: nods thoughrfgully
  • [14:05] Zha Ewry: That just requires stashing the association
  • [14:05] Zha Ewry: we end up with a lot of stashed state for the logged in ave, but that's sort of unavoidable
  • [14:06] Tess Linden: Zha: what do you mean stashed state?
  • [14:06] Teravus Ousley: nods.. gotta pass them around in order to get them back to vaak and other LL systems
  • [14:06] Zha Ewry: You need to keep some state about whence the ave came from, and what cap you issued, and such. its pretty unavoidable
  • [14:07] Zha Ewry: The only place it gets tricky, is deciding when you can flush it
  • [14:07] Zha Ewry: You want to keep it until the last moment, on the way out, so you can handle timed out, and failed derez requests
  • [14:08] Zha Ewry: I don't think its avoidable, given the nature of doing this REST style
  • [14:08] Tess Linden: the way region seed caps currently work is if you stop polling the event queue it'll revoke your caps
  • [14:08] Teravus Ousley: well, I think it's clear that this is primarily a grid service. So, regardless of which region a user is in.. the state should be kept until the avatar moves to a different region domain.
  • [14:08] Teravus Ousley: .. as far as the region domain is concerned
  • [14:08] Zha Ewry: yes
  • [14:09] Tess Linden: it can be tricky though. On agni today, you can have the seed cap of a neighboring region in order to get updates
  • [14:09] Saijanai Kuhn: you need to pol the region cap?
  • [14:09] Tess Linden: you poll the event queue cap thats granted by the region cap
  • [14:09] Zha Ewry: I just noticed some remarkably careful to watch holes you could get on not quite right derezes
  • [14:10] Saijanai Kuhn: that's different than the cap for AD presence?
  • [14:10] Whump Linden: I think there's a confusion over the region seed cap and if its something kept around througout the agent's session connected to the agent domain?
  • [14:10] Teravus Ousley: notes that the event queue isn't implemented on OpenSim.. so it's a moot point as far as that's concerned.. but could just as easily revoke it on derez
  • [14:11] Zha Ewry: yeah, tho, i expect, sooner or later, probably sooner, we'll have to impement it
  • [14:11] Teravus Ousley: agreed.. I've brought it up several times..
  • [14:11] Zha Ewry: nods
  • [14:11] Zha Ewry: We should take tha offline, We need to chase some stuff in terms of doing that in a good performing fashion
  • [14:12] Saijanai Kuhn: EQG for agent Domain seems to take forever to return sometimes
  • [14:12] Tess Linden: whats nice about llidl is it looks a lot like code, which includes error responses
  • [14:12] Zha Ewry: nods
  • [14:12] Teravus Ousley: Yes, it's much more clear described that way, Tess
  • [14:12] Zha Ewry: I would have loved to see that when I was doing the POC coding ;-)
  • [14:13] Tess Linden: Each of these notes here should be in a Jira task for getting the grid up to spec
  • [14:13] Tess Linden: well, our code isn't even close to the spec yet, so theres a lot more to code against this new spec
  • [14:14] Tess Linden: Whump? Im ready to pass on the mic
  • [14:14] eaglefx Binder: Good Evening
  • [14:14] Whump Linden: Okay, thanks Tess, Infinity.
  • [14:14] Infinity Linden: yer welcome... just happy the specs escaped
  • [14:14] Xugu Madison: Indeed, thanks Tess & Infinity!
  • [14:14] Tess Linden:  :) me too, thanks for bringing us in whump
  • [14:15] Whump Linden: So, I'll have a schedule for how we are going to make Draft 3 live within the next few days.
  • [14:15] Zha Ewry: now has this image of the specs sneakign around on little legs leaping onto the wiki when Infinity wasn't looking.
  • [14:15] Infinity Linden: and Whump
  • [14:15] Infinity Linden: i notice that Tao's not here
  • [14:15] Whump Linden: So other updates, we are down to six open issues on Jira.
  • [14:15] Infinity Linden: he's probably going to have something to say about the deploy as well
  • [14:15] Saijanai Kuhn: vacation for 2 weeks. I'm his not-so-competent proxy
  • [14:15] Teravus Ousley: Hmm.. ok so the same cap will be invoked on derez.. we'll keep state on the avatar, so will know if it's being invoked based on our state for rez or derez
  • [14:15] Infinity Linden: okay
  • [14:15] Infinity Linden: forgot about that
  • [14:16] Whump Linden: Oh, Teravus, you had a question yesterday from OpenSim OO about the UUID we pass.
  • [14:16] Zha Ewry: listens
  • [14:17] Whump Linden: " however it's been requested that the client attempt to connect to the region with a UUID given to it by the OGP service instead of the Aditi UUID."
  • [14:17] Teravus Ousley: oh, yes.. I was just asking if there was a way that we can tell the agent domain to have the client use a specific UUID in the UseCircuitCode packet
  • [14:18] Teravus Ousley: .. that way, we could ensure.. that there was no way for someone to manually mangle their avatar/UUID in the database and have them overlap
  • [14:18] Zha Ewry: This is going to get us into the messy is it fname/lname ot uuid that matters discussion
  • [14:18] Infinity Linden: blergh.
  • [14:18] Teravus Ousley: hehe
  • [14:18] Infinity Linden: we should bring it up with Zero tomorrow morning
  • [14:19] Zha Ewry: (and it probably, long term needs to be fname/lname/domain, or uuid/domain, anyway
  • [14:19] Infinity Linden: identity registration has been getting a lot of discussion lately
  • [14:19] Infinity Linden: +1 Zha
  • [14:19] Zha Ewry: since, there's no coherent way to manae fname/lname == uuid globally without going insane
  • [14:19] Infinity Linden: funny... i have a note on my desk right now that says exactly the same thing
  • [14:19] Zha Ewry: *MANAGE
  • [14:19] Teravus Ousley: :D
  • [14:19] Infinity Linden: not MANGLE?
  • [14:20] Zha Ewry: Oh, mangling is easy
  • [14:20] Zha Ewry: Managing it coherently, that's harder
  • [14:20] Teravus Ousley: haha, it is indeed
  • [14:21] Zha Ewry: Wow. We've been driven into silence
  • [14:22] Whump Linden: Okay, any other questions?
  • [14:22] Infinity Linden: 'cause we're so 4w3sum?
  • [14:22] Teravus Ousley: shakes head. LLIDL = clear
  • [14:22] Zha Ewry: is power skimming the printed documents
  • [14:22] Whump Linden: There are some OSGrid regions running OGP as of now.
  • [14:22] Xugu Madison: Whump, very quickly before we wrapup; totally failed to do anything I said I'd do, sorry, someone put a digger through the powerline at work and it all went downhill from there
  • [14:23] Zha Ewry: Out of curisoity, the sequence diagrams while not definitive, are we goign to look at having them in the spec, in time? They're really nice to be able to skim
  • [14:23] Infinity Linden: WRT LLIDL and Auth... i'll try to be more available over the next week for discussions in world or via email
  • [14:23] Whump Linden: Xugu no problem. Backhoes do that sort of thing all the time.
  • [14:24] Infinity Linden: lol... Zha... Tess, Whump and I were having that exact discussion yesterday
  • [14:24] Zha Ewry: chuckles
  • [14:24] Zha Ewry: did you come to any conclusions?
  • [14:24] Zha Ewry: i'd vote for them in
  • [14:24] Infinity Linden: i think we agreed that our preference is for the specs to be mostly textual
  • [14:24] Tess Linden: Zha: zero says that the UML diagrams may contradict the spec so cannot be in the spec itself but should maybe be part of a "Explained" doc
  • [14:24] Teravus Ousley: hurricanes also delay flights.. even 5 days in advance
  • [14:25] Zha Ewry: thinks
  • [14:25] Zha Ewry: that requires one more bit of synced paper
  • [14:25] Infinity Linden: might be a good topic for Zero's OO's tomorrow morning
  • [14:25] Zha Ewry: I'm perfectly happy with them being tagged as "non normative"
  • [14:25] Infinity Linden: ah
  • [14:25] Infinity Linden: okay
  • [14:25] Zha Ewry: Good topic for tommorow
  • [14:25] Infinity Linden: we're very interested in having a "normative" spec
  • [14:26] Infinity Linden: and a separate exegesis document
  • [14:26] Zha Ewry: nods
  • [14:26] Infinity Linden: both of which are "official"
  • [14:26] Infinity Linden: but the spec defines mechanism, not policy
  • [14:26] Zha Ewry: That always leaves the sync issue, but either approach is valid
  • [14:27] Infinity Linden: and the "OGP Explained" doc privides more guidance
  • [14:27] Infinity Linden: yeah... i think we're paranoid that ew might have a normative spec where diagrams contradict the text
  • [14:27] Whump Linden: I've updated that with the new teleport UML
  • [14:27] Infinity Linden: it would be "very bad"
  • [14:27] Zha Ewry: nods
  • [14:27] Whump Linden: https://wiki.secondlife.com/wiki/OGP_Explained#OGP_Draft_3_Teleport
  • [14:28] Zha Ewry: well, all you ve to do, is tag the diagrams as non-normative
  • [14:28] Infinity Linden: and if the OGP Explained is "not exactly normitive" it's less of a big deal
  • [14:28] Zha Ewry: btu.. I have to say, I've been on both sides of this debate, over the years
  • [14:28] Infinity Linden: hmm... well.. the specs will eventually be published in XML
  • [14:28] Infinity Linden: you could break out the XSLT foo and create a doc that combines the two
  • [14:28] Zha Ewry: wimpers
  • [14:29] Zha Ewry: and never know if you were mungign synced versions
  • [14:29] Infinity Linden: XML Heretic!!!!
  • [14:29] Zha Ewry: dimly recalls an OMG meeting in philadelphia and looks for something to drink
  • [14:29] Whump Linden: okay, and with that....
  • [14:29] Infinity Linden: the use of XML ensures "compatibility with the future"
  • [14:29] Infinity Linden: (that was a joke, btw)
  • [14:29] Teravus Ousley: haha
  • [14:30] Zha Ewry: Okies
  • [14:30] Infinity Linden: lol
  • [14:30] Zha Ewry: Very nice, and glad it's out
  • [14:30] Infinity Linden: I was Borland's rep to the OMG for a couple years
  • [14:30] Zha Ewry: Now to read and comment, and such
  • [14:30] Infinity Linden: after VisiJeff left
  • [14:30] Whump Linden: thinks maybe I should put in a bar, or a least a stash of scotch.
  • [14:30] Infinity Linden: i definitely remember that most of OMG spec foo was decided at the bar
  • [14:31] Zha Ewry: is deeply suspciious we shared OMG meetings
  • [14:31] Teravus Ousley: hehe
  • [14:31] Infinity Linden: probably
  • [14:31] Infinity Linden: don't remember philly
  • [14:31] Infinity Linden: but sevral in boston and at least one in Mesa, AZ
  • [14:31] Infinity Linden: oh oh
  • [14:31] Infinity Linden: alos
  • [14:31] Infinity Linden: also
  • [14:31] Zha Ewry: recalls Philly, Denver, LA, Boston, Mesa, London, Tokyo
  • [14:32] Infinity Linden: someone mentioned that the architecture of X-Windows was eventually finalized on a napkin at Jing Jing in Palo Alto
  • [14:32] Teravus Ousley: yay for napkin notes..
  • [14:32] Infinity Linden: we're considering taking a field trip there if for no other reason than to get that g33k mojo working for us
  • [14:32] Zha Ewry: laughs
  • [14:32] Teravus Ousley: has a few of those framed himself
  • [14:32] Zha Ewry: really
  • [14:33] Zha Ewry: at some point, I should tell the horrible story of the first DSOM server deploy I ever did.
  • [14:33] Infinity Linden: remembers the days when business plans could be submitted to VCs on bar napkins
  • [14:33] Infinity Linden: 94?
  • [14:33] Zha Ewry: 93
  • [14:33] Zha Ewry: internal alpha
  • [14:33] Infinity Linden: I was working with the toro-lab folk hten
  • [14:33] Infinity Linden: having just migrated from Convex
  • [14:34] Zha Ewry: I will merely wince and observe that the orginal alpha code, put a class on one and only one ORB
  • [14:34] Xugu Madison: I have to run, thanks for hosting Whump, see you later all!
  • [14:34] Infinity Linden: don't tell me... let me guess... licence manager problems?
  • [14:34] Whump Linden: thanks's Xugu
  • [14:34] Zha Ewry: and if you had a class which laminated several classes, and they ended up on
  • [14:34] Zha Ewry: multiple ORBS
  • [14:34] Zha Ewry: you could have several dozen
  • [14:34] Whump Linden: I think we're done with the meeting and into the swapping of war stories.
  • [14:34] Zha Ewry: RPCs to construct a class
  • [14:34] Infinity Linden: yay! and people wonder why we don't just adopt CORBA as our communication middleware!
  • [14:34] Mirt Tenk: ty, srry late
  • [14:35] Zha Ewry: yeah
  • [14:35] Infinity Linden: totally
  • [14:35] Zha Ewry: So, We have a ocuple of topics for tomorow's office hours. Yeah!
  • [14:35] Infinity Linden: reviewing the scars we recieved after falling off bar stools at various standards meetings
  • [14:35] Infinity Linden: yup
  • [14:35] Zha Ewry: Super
  • [14:36] Infinity Linden: i'll ping zero so it's not a surprise
  • [14:36] Teravus Ousley: :D
  • [14:36] Infinity Linden: oh wait
  • [14:36] Whump Linden: Infinity: Zero is in transit.
  • [14:36] Infinity Linden: he's working out of a different office
  • [14:36] Zha Ewry: Then we can start wrassling about how we want to do cross component proof of truyst in the large
  • [14:36] Infinity Linden: yup
  • [14:36] Infinity Linden: i don't think he's coming in to our office before going home
  • [14:36] Infinity Linden: maybe an email
  • [14:36] Zha Ewry: ok, everyone see you all anon
  • [14:37] Infinity Linden: cheers
  • [14:37] pjay Luik: please help me
  • [14:37] Teravus Ousley: ok, thanks for coming.. have a great evening.. afternoon
  • [14:37] Whump Linden: yes, thanks everyone
  • [14:37] Infinity Linden: i'm going to escape so i can get something done
  • [14:37] Cane Janick: Thank you
  • [14:37] Whump Linden: woah, I see Nebadon and hear a Yoko Kanno soundtrack.
  • [14:38] Nebadon Izumi: hehe
  • [14:38] eaglefx Binder:  ;)
  • [14:38] Nebadon Izumi: yea im tired of being human
  • [14:38] pjay Luik: what
  • [14:38] eaglefx Binder: cool robot
  • [14:38] Teravus Ousley: hy00man?
  • [14:38] Nebadon Izumi: i'll try not to step on anyone
  • [14:38] Nebadon Izumi: :D
  • [14:39] eaglefx Binder: hehe
  • [14:39] pjay Luik: where can i find it
  • [14:39] Nebadon Izumi: this robot?