User:Zero Linden/Office Hours/2007 Nov 13

From Second Life Wiki
Jump to: navigation, search
  • [13:01] Rex Cronon: it seems that both zero and rob landed in the car:)
  • [13:01] Zero Linden: what car? :-)
  • [13:01] Rex Cronon: hello zero
  • [13:01] Zero Linden: hello all
  • [13:02] Morgaine Dinova: 'Afternoon Zero :-)
  • [13:02] Rob Linden: hi all
  • [13:02] Masaw Umaga: hi!
  • [13:02] Rex Cronon: there was a car in the same place u rezzed:)
  • [13:02] Rex Cronon: hi rob
  • [13:02] Zero Linden: I know... I returned it
  • [13:02] Zero Linden: we'll wait a few more minutes for folks
  • [13:03] Lily Argus: Folks is here, we were just waiting for you. .P
  • [13:03] Saijanai Kuhn: its the day after a holiday. no-one was at theGroupies meeting either
  • [13:03] Morgaine Dinova: I'm beginning to think that attendence figures on a Tuesday are random
  • [13:04] Zero Linden: I'm pretty sure Burhop wanted to be here
  • [13:04] Morgaine Dinova: He was at Groupies
  • [13:05] Zero Linden: but not here - yet?
  • [13:05] Zero Linden: well, let's get on with it!
  • [13:05] Zero Linden: Welcome to my office hours
  • [13:06] Morgaine Dinova: Oh Zero, I only noticed yesterday a question you asked about proxy caches in Brainstorming. I answered it. Sorry didn't notice before.
  • [13:06] Zero Linden: As you may or may not know - these are held in public, and the transcript is posted to the wiki
  • [13:06] Zero Linden: Speak your mind, but speak to the world!
  • [13:07] Tao Takashi: Hi
  • [13:07] Zero Linden: Now - first I have a procedural question to ask
  • [13:07] Zero Linden: and it is a bit of a favor
  • [13:07] Saijanai Kuhn: hey tao, as expected your code didn't quite work ;-)
  • [13:07] Rex Cronon: does anybody know how many moded viewers are available?
  • [13:07] Tao Takashi: Sai: I guessed so ;-)
  • [13:07] Tao Takashi: SL changes to fast ;-)
  • [13:08] Morgaine Dinova: Tao: it's the Singularity ;-)
  • [13:08] Zero Linden: I would like to move the Thursday meetings
  • [13:08] Zero Linden: Either one hour later, to 8:30am
  • [13:08] Zero Linden: OR
  • [13:08] Zero Linden: to Wednesday (keeping it at 7:30am)
  • [13:08] Zero Linden: any opinions?
  • [13:09] Rob Linden: Rex, Im guessing there are about half a dozen that get above the 100 download/month mark
  • [13:09] Rex Cronon: the same time as andrew?
  • [13:09] Morgaine Dinova: Tues+Weds sound a bit too close to each other. But I'm fine with either change.
  • [13:09] Rex Cronon: thanks rob
  • [13:09] Dizzy Banjo: greetings
  • [13:09] Tao Takashi: I don't care
  • [13:09] Rex Cronon: hi
  • [13:09] Tao Takashi: whatever is best for the rest :)
  • [13:09] Morgaine Dinova: Hi Rob
  • [13:09] Saijanai Kuhn: I seldome go to andrew's meeting but a change to avoid conflict is nice
  • [13:09] Finrod Meriman: both work
  • [13:09] Savannah Glimmer: Hi Dizzy
  • [13:10] Morgaine Dinova: Hi Dizzy
  • [13:10] Dizzy Banjo: howdy
  • [13:10] Zero Linden: okay - this Thursday willbe the same, and I'll ask there as well, but I'll register a distinct "either way" from this crowd
  • [13:11] Rex Cronon: zero, could u try and ask this question on the main blog, and see how may people bother to answer it?
  • [13:11] Zero Linden: Possibly - I hate to put something like that on the main blog......
  • [13:11] KK Jewell: hello
  • [13:11] Eshi Otawara: winks at KK
  • [13:12] KK Jewell: No room?
  • [13:12] Zero Linden: it gets in the middle of the feed for most residents
  • [13:12] Zero Linden: oh well,
  • [13:12] Zero Linden: moving on
  • [13:12] Zero Linden: So, no Burhop, eh?
  • [13:12] Zero Linden: he wanted to discuss geomtery
  • [13:12] Morgaine Dinova: Did he have an agenda point?
  • [13:12] Zero Linden: should we have that conversation without him, or just table it?
  • [13:12] Morgaine Dinova: Aha
  • [13:13] Morgaine Dinova: Table it unless some other main player from that VAG here. I'm only an observer there.
  • [13:13] Zero Linden: he's not online
  • [13:13] Zero Linden: three...
  • [13:13] Zero Linden: two...
  • [13:13] Zero Linden: one...
  • [13:13] Zero Linden: tabled!
  • [13:13] Morgaine Dinova:  :P
  • [13:13] KK Jewell: Iota I am over here
  • [13:13] Zero Linden: So - last time we had a really good conversation about
  • [13:13] Zero Linden: how getting inventory would work
  • [13:14] Iota Ultsch: Hello
  • [13:14] Zero Linden: did anyone have any questions about that, after it has sat for a few days?
  • [13:14] Tao Takashi: I actually need to read up on it ;-)
  • [13:14] Rex Cronon: sorry, but what do u mean by table it? u intendt to use it at a later time?
  • [13:14] Saijanai Kuhn: I've got a simpler question, but probably can wait
  • [13:14] Morgaine Dinova: Rex: yes, deferred
  • [13:14] Rex Cronon: ok, thanks
  • [13:15] Zha Ewry: So, Zero, did you capture all the idemopttent/rest discussion?
  • [13:15] Tao Takashi: would be more interested what the state of making the wiki faster is ;-)
  • [13:15] Zero Linden: Zha - I understood that Which was going to do that
  • [13:15] Zero Linden: and he scheduled another hour with me about it for tomorrow
  • [13:15] Morgaine Dinova: Tao: seconded. And since Rob is here, we can ask that :-)
  • [13:16] Zha Ewry: Yesterday
  • [13:16] Zero Linden: Rob - is there any news?
  • [13:17] Zero Linden: hmmm.... suppose not...
  • [13:17] Rob Linden: sorry, got distracted
  • [13:17] Zero Linden: Ah
  • [13:17] Rob Linden: I'm hopeful that some changes in the next couple of weeks will help things
  • [13:18] Morgaine Dinova:  ????
  • [13:18] Tao Takashi: thinks about mirroring it so there's at least a read-only copy ;-)
  • [13:18] Rob Linden: there's some hardware upgrades and configuration changes slated in a couple of weeks
  • [13:18] Rob Linden: ...on wiki.secondlife.com
  • [13:18] Morgaine Dinova: So there is no serious attempt ongoing at fixing the wiki then. kk
  • [13:18] Rob Linden:  ?
  • [13:19] Morgaine Dinova: Rob, we've had the same answer for a couple of months.
  • [13:19] Tao Takashi: just reading the wiki: "asking the seed capability" on an agent host means POST, right?
  • [13:19] Tao Takashi: e.g. for getting caps
  • [13:20] Rob Linden: feel free to host your own Morgaine
  • [13:20] Morgaine Dinova: The community is trying to help you Rob, not vice versa.
  • [13:21] Lalinda Lovell: damn here she goes
  • [13:21] Zero Linden: Yes, Tao, It does, it means POST to the seed-cap, with a body that lists the capabilities you'd like granted
  • [13:21] Tao Takashi: ok
  • [13:21] Tao Takashi: we just discussed login in the AWGroupies meeting and I wasn't aware of what was said in the last OH
  • [13:21] Tao Takashi: so it answers some questions but basically we came to similar conclusions ;-)
  • [13:22] Zero Linden: There are some details there to be figured out - like do you have to present the feature sets you want at the initial log in-
  • [13:22] Zero Linden: or do we just reserve that until you ask for them from the seed-cap
  • [13:22] Tao Takashi: of course I don't expect it to be final ;-)
  • [13:22] Zha Ewry: Thge later, is easier
  • [13:22] Saijanai Kuhn: yeah, I'm trying tor udnerstand the current login A lot of "extraneous" stuff seems to be required currently
  • [13:23] Tao Takashi: actually I am all for an iterative process. I'd even start without any caps and add them later, just to get some sort of idea for the flow
  • [13:23] Zero Linden: I like the former becuase then the agent domain can decide to issue you a seed-cap to a less resource intensive service
  • [13:23] Zha Ewry: Becuase the more you pass in the login, the more you have to deal with wehen ou may not need it at all
  • [13:23] Zero Linden: if all you want is to, say, do IM with your friends
  • [13:23] Saijanai Kuhn: do we need ANYTHING, beyond name/password?
  • [13:23] Tao Takashi: maybe you just need an openid ;-)
  • [13:23] Zero Linden: Well, name/credentials (not all Agent domains may use passwords, one expects...)
  • [13:23] Saijanai Kuhn: sure.
  • [13:24] Saijanai Kuhn: The current method seems to ahve some "optional" stuff that is required, but maybe its not
  • [13:24] Tao Takashi: it should be pluggable of course but I'd think that name/creds is all what is needed
  • [13:24] Tao Takashi: if you want to login to a certain region you might give this information later
  • [13:24] Tao Takashi: what else is transferred actually these days?
  • [13:24] Zero Linden: The credential allows an agent domain to use OpenID - which would have to be via a web browser (OpenID relies on being able to redirect the browser to show the user a page)
  • [13:24] Tao Takashi: I only have old code ;-)
  • [13:25] Zero Linden: and then use the result of that interactionl, a one-time credential, to login in
  • [13:25] Rob Linden: has half-composed email to sldev about openid, oauth, and Google's authsub
  • [13:25] Tao Takashi: Zero: I'd expect to have browser and SL client in one thing anyway ;-) for IM even a browser with some extension might be enough
  • [13:25] Tao Takashi: I just met David Recordon in Berlin, btw
  • [13:26] Tao Takashi: he gave a talk on OAuth
  • [13:26] Zero Linden: What I'm wondering is if there is some set of uses of logging-in that should induce an allocation of a live process on an Agent Host
  • [13:26] Tao Takashi: and OpenSocial and OpenID
  • [13:26] Saijanai Kuhn: right now, libsl and the client pass about a dozen parameters and these appear to gqueue stuff on the server side
  • [13:26] Zero Linden: "Open" is the new "X"
  • [13:26] Saijanai Kuhn: seems like we need to decide what we don't need during the initial login.
  • [13:26] Rob Linden: (AuthSub is the authentication component of OpenSocial)
  • [13:26] Tao Takashi: I would think that OAuth might be handy for linking avatars or something like that, basically cross-domain authorization
  • [13:27] Tao Takashi: Rob: For now, I think there's some plan or at least rumour to switch to OAuth
  • [13:27] Zero Linden: I'm not sure, I've read through the OAuth spec once....
  • [13:27] Tao Takashi: ok, I am not sure either ;-)
  • [13:27] Zero Linden: ....so far it reads a little too open-ended tome....
  • [13:27] Saijanai Kuhn: we do need name/cred. ANything else, or can that be requested later?
  • [13:27] Zero Linden: ...but I concievably haven't groked it yet.
  • [13:27] Tao Takashi: but we might have a closer look later on
  • [13:28] Tao Takashi: actually I gave a Barcamp session about OpenSocial with David, that was quite a nice discussion
  • [13:28] Zero Linden: I suppose we could put the client-version info into the HTTP headers as user-agent
  • [13:28] Zero Linden: so, probably nothing else
  • [13:28] Tao Takashi: [(http://tinyurl.com/2ebzgs)]
  • [13:28] Saijanai Kuhn: hmmmm...
  • [13:29] Saijanai Kuhn: so would the wrong client trigger an error, or a fallback method for communicating?
  • [13:29] Rob Linden: is still trying to grok all of the different auth proposals out there, and hoping discussion here and on sldev can hlep
  • [13:29] Zero Linden: And the result is a set of capabilities - seed-cap is required, but we should allow login to return other caps - unasked for
  • [13:29] Zha Ewry: As long as they are fully self desctribed, sure
  • [13:30] Zero Linden: I'm thinking "Ah, I see you are logging in from a full-3D viewer, I *know* you'll want IM, and chat, and inventory and group capabilities"
  • [13:30] Zero Linden: caps are currently named via a string
  • [13:30] Tao Takashi: Rob: which different ones do you mean? openid, oauth and authsub?
  • [13:30] Zero Linden: though it should probably be a qualified name tree: sl.inventory.getroot
  • [13:30] Rex Cronon: shouldn't the viewer ask for these things?
  • [13:30] Zero Linden: so that we can have things like: x.example.sky.cloud-control
  • [13:30] Rex Cronon: why does the server assume the viewer wants them?
  • [13:31] Saijanai Kuhn: and does an IM client need sl.inventory.getroot?
  • [13:31] Morgaine Dinova: There are likely to be so many different cliente that the server had better not infer anything about required caps. Let the client ask for them.
  • [13:31] Zero Linden: Rex, sure, but since the common case is *SO* common, we could allow login to return extra things as an optimization
  • [13:31] Tao Takashi: does the explanation of inventory stop with 3) get-root and get-contents? :)
  • [13:31] Rex Cronon: right. let the client request what it wants
  • [13:31] Zero Linden: reducing latency on log-in is important to user experience
  • [13:32] Rob Linden: Tao: ^^ yeah (/me tries to avoid taking us on tangent)
  • [13:32] Zha Ewry: I think, ideally, ways to cue, "Give me X" and "don't give my Y" on login may be the way to go there, then
  • [13:32] Zero Linden: Tao - it is a little complicated by "category" folders and we haven't discussed moving them or creating them
  • [13:32] Tao Takashi: but isn't openid something different than e.g. oauth and authsub?
  • [13:32] Saijanai Kuhn: could we have a parameter on login that suggests that optional stuff not needed? IM clients don't need much of anything
  • [13:32] Tao Takashi: aren't the other two just used for doing cross-website authorization a la flickr?
  • [13:33] Zero Linden: Tao - openid is a way to allow a user to prove to a "relying party" that they have authorized control over a particular URL
  • [13:33] Saijanai Kuhn: so to reduce server load and to reduce lag, we could bypass almost all "optional" stuff
  • [13:33] Zero Linden: I'm not sure how the auth proposals link into them - to be honest
  • [13:33] Tao Takashi: which URL do you mean? the openid itself?
  • [13:34] Zero Linden: Sai - we should reduce it in places where we know the latency is a bear
  • [13:34] Tao Takashi: well, I think that once you are identified on service A by your openid and you want to access something from service B then OAuth comes into play as service A asks service B if that's ok
  • [13:34] Zero Linden: but perhaps, yes, we could just go minimal in the first round here:
  • [13:34] Morgaine Dinova: Just decouple all the caps as much as possible, and let clients select. The decoupling will pay dividends for short-term scalability too ... disable the bits that are lagging you.
  • [13:34] Zero Linden: login -> seed -> other caps as requested
  • [13:34] Rob Linden: Tao: I suspect that a lot of people look at OpenID right now when all they need is something like OAuth. However, i understand that they are supposed to be somewhat complementary, but I'm still grokking how
  • [13:35] Saijanai Kuhn: perfect
  • [13:35] Zha Ewry: Yeah, exactly, Zero.
  • [13:35] Zha Ewry: Simple and fast and works well for all the clients
  • [13:35] Tao Takashi: and as OAuth seems a generalized solution to AuthSub and the thing Yahoo uses I think we should go with Oauth
  • [13:35] Tao Takashi: as I understand it the goal here is to make all the parties switch to OAuth
  • [13:35] Rob Linden: Yahoo's thing: bbAuth
  • [13:35] Tao Takashi: that's how I understood David
  • [13:35] Zero Linden: I'm not sure where we'd use OAuth
  • [13:35] Zero Linden: Are you?
  • [13:35] Tao Takashi: as they all got together but nobody wanted to use their competitors tools, thus created a new one
  • [13:36] Tao Takashi: for now I am also not sure :)
  • [13:36] Tao Takashi: I first need to see the cross domain thing
  • [13:36] Tao Takashi: what actually is needed there
  • [13:36] Tao Takashi: how trust is created, how e.g. links between agents can be made
  • [13:36] Tao Takashi: when we are in this discussion I guess we can decide that better
  • [13:37] Tao Takashi: so from what I understood I also think that openid and oauth are complementary
  • [13:37] Rob Linden: OpenID is for when you want anyone to be an identity provider. OAuth is when you still want central identity but decentralized services. OAuth may be a good baby step to OpenID, if I understand things correctly.
  • [13:37] Tao Takashi: you login to both service A and B with some openid (maybe the same) and then it's just the question how to allow service A to access something from service B
  • [13:37] Zero Linden: I'm not sure I read it that way
  • [13:38] Zero Linden: well.... obviously we all need to go read some specs....
  • [13:38] Tao Takashi: I mean OAuth is basically what flickr does when you want to access it from some e..g photo printing service
  • [13:38] Tao Takashi: this is also what some example shows
  • [13:38] Tao Takashi: I will find it, one moment
  • [13:38] Rob Linden: Zero: yup
  • [13:38] Rob Linden: (the result I was hoping for)
  • [13:38] Rob Linden:  :)
  • [13:39] Tao Takashi: honestly I also need to read the spec. I am more or less quoting from David's barcamp session ,-)
  • [13:39] Zero Linden: But then, it amounts to capabilities, does it not?
  • [13:40] Tao Takashi: well, sort of I think
  • [13:40] Tree Kyomoon: can someone tell me the difference between "capabilities" and "webservices" ?
  • [13:40] Tao Takashi: [1]
  • [13:40] Zero Linden: hmmmm... okay - homework for us all
  • [13:40] Tao Takashi: this is the demo
  • [13:40] Tree Kyomoon: (dumb question sorry)
  • [13:40] Tao Takashi: basically looks like flickr
  • [13:41] Gareth Ellison: may i ask which spec people are talking about here? the general OpenID specs or something more specific?
  • [13:41] Zero Linden: Tree - a capability is a non-guessable URL that lets anyone in it's posession invoke a service
  • [13:41] Tao Takashi: hey, you have all the people in town in SF ;-)
  • [13:42] Tao Takashi: just email David and ask him, after all he is co-inventor of it ;-)
  • [13:42] Tree Kyomoon: but not a hashed URL, just a weird long one?
  • [13:42] Tao Takashi: so the page says, OAuth is for API access delegation
  • [13:42] Zero Linden: A web service, as usually deployed, require the user to "authenticate" on each use - often by supplying an id and password or by signing the request
  • [13:42] Gareth Ellison: Tree: UUIDs are used
  • [13:42] Saijanai Kuhn: is it like a series of bit flags or just a randome string?
  • [13:42] Zero Linden: so each function must also authenticate
  • [13:42] Tree Kyomoon: webservices usually use sessions (cookies) i think
  • [13:42] Zero Linden: capabilities separate those functions - you authenticate to one service, who authorizes you to do somethin by giving you the capability
  • [13:43] Zero Linden: using the capability doesn't requires authentication or authorization
  • [13:43] Saijanai Kuhn: so that string is an authorization for one or more functions?
  • [13:43] Finrod Meriman: and it assumes a secure (unsnoopable) channel.. yes?
  • [13:43] Zero Linden: Sai - exactly
  • [13:43] Gareth Ellison: Saijanai: read your viewer debug logs
  • [13:43] Tao Takashi: so in fact it's a baked in authorization, isn't it?
  • [13:43] Gareth Ellison: when you cross a sim border
  • [13:44] Gareth Ellison: you see the viewer sending a POST to a CAP
  • [13:44] Zero Linden: there are two advantages to capabilities over normal web services
  • [13:44] Tree Kyomoon: so capabilities dont use sessions, so you cant really time a person out?
  • [13:44] Saijanai Kuhn: might be built into how long the cap remains viable
  • [13:44] Saijanai Kuhn: use or lose after x minutes
  • [13:44] Zero Linden: 1) on the back end, the code for authentication and authorization needn't be invoked from every service entry point.....
  • [13:44] Tao Takashi: the server can do a timeout I guess
  • [13:44] Zero Linden: it is centralized in one place
  • [13:44] Zha Ewry: You can always go to leasing models
  • [13:44] Tao Takashi: after a while it's not valid anymore
  • [13:45] Zero Linden: 2) You can give the capability to something else to use it without giving them full access to your account
  • [13:45] Tao Takashi: you might even automate this by baking in some time token if the issuer wants to say that it's only alive for x minutes
  • [13:45] Tao Takashi: so in fact 2) sounds like OAuth ;-)
  • [13:45] Zero Linden: Consider when Facebook says that they'll search your AIM account for friends that are in Facebook
  • [13:45] Zero Linden: you have to give Facebook your aim account name and password
  • [13:45] Zero Linden: so they can invoke the web service "get friends list"
  • [13:45] Saijanai Kuhn: that could still be part of hte backend. Gimme a cap that is good for x minutes for y ad z services
  • [13:46] Zero Linden: since they have to authenticate that operation
  • [13:46] Zero Linden: but instead, if the user went to AIM and said - give me a URL to give Facebook so they can get my friends list exactly once
  • [13:46] Zero Linden: then that URL would be a capability
  • [13:46] Tree Kyomoon: so a cap basically does what sessions do, but puts the session rigth into the URL?
  • [13:46] Zha Ewry: And makes it more secure
  • [13:46] Zero Linden: as you give that URL to Facebook, and when they invoke it (POST or GET)
  • [13:46] Zero Linden: it returns the friends list
  • [13:47] Tree Kyomoon: ah ok I see the difference then
  • [13:47] Zero Linden: Tree - yes, but they are more fine grained than sessions
  • [13:47] Zero Linden: since typically sessions just say you are you and therefore have full access
  • [13:47] Zero Linden: the cap is only good for one specific operation
  • [13:47] Gareth Ellison: a session is longer-lasting and grants permissions for all operations
  • [13:47] Gareth Ellison: yeah - single operation
  • [13:47] Tree Kyomoon: so your capability server has to be pretty scookum!
  • [13:47] Gareth Ellison: should get out of the habit of talking over Zero
  • [13:47] Zero Linden: "scookum"?
  • [13:48] Gareth Ellison: tree: scookum == high capacity?
  • [13:48] Tree Kyomoon: powerful/robust
  • [13:48] Zero Linden: it is about 500 lines of python, writen on eventlet and mulib
  • [13:48] Tao Takashi: actually I am just wondering right now how I would implement such a caps server ;-)
  • [13:48] Zero Linden: it is pretty simple straight forward code
  • [13:48] Saijanai Kuhn: but it can be a combination of services, right? I can request data in a stream and it can send me 10 bits of data without reqring a new cap
  • [13:48] Tree Kyomoon: yes, it sounds like a very clever approach
  • [13:48] Gareth Ellison: mulib makes it a simple dict operation
  • [13:48] Tao Takashi: ok, maybe a simple mapping
  • [13:48] Zha Ewry: Look at hydra/mmp for ancient history
  • [13:48] Zero Linden: Saijanai - it is just like API design - you have to come up with how much info is the cap for
  • [13:49] Gareth Ellison: on the current grid you get sent a CAPS seed at login
  • [13:49] Gareth Ellison: or on border crossings
  • [13:49] Zero Linden: is the cap for "get parcel info" where the parcel id is in the request body
  • [13:49] Saijanai Kuhn: right, so it could be one thing or a shopping list, as agreed upon in the design
  • [13:49] Tao Takashi: would once again use Python interfaces and utilities for registering new capabilities
  • [13:49] Zero Linden: or is the cap for "get parcel info for parcel id 1234456"
  • [13:49] Zha Ewry: And.. the other thing, I think, is that, with these beign fine grained, we should be able to change them out, as we learn, if we guess wrong
  • [13:49] Zero Linden: Tao - actually, there is a web service for granting and revoking capabilities
  • [13:49] Zero Linden:  !
  • [13:50] Gareth Ellison: Zero: i'd argue that you should allow fine-grained CAPs to be built by saying "send me a more limited version of my more generic CAP"
  • [13:50] Zero Linden: so, the fact that it is in python is implemetnation only
  • [13:50] Zero Linden: Gareth- a sort of "bind" operation
  • [13:50] Tao Takashi: btw, is there some LLSD lib for python somewhere? Probably I asked this already
  • [13:50] Zero Linden: Tao - yes,
  • [13:50] Zero Linden: I don't know if it is in the open source stuff yet or not, but it shouldl be
  • [13:50] Zero Linden: Rob, do you know?
  • [13:51] Gareth Ellison: there is
  • [13:51] Tao Takashi: it should be so I can start coding the actual protocol and not first start with some stuff around it ;-)
  • [13:51] Rob Linden: looks up url
  • [13:51] Zero Linden: Tao, you don't like Yak shaving?
  • [13:51] Saijanai Kuhn: Tao, can I help with that?
  • [13:51] Gareth Ellison: wonders if anyone looked at his abandoned AWG SVN
  • [13:51] Saijanai Kuhn: Would be first part of modular client VAG
  • [13:52] Saijanai Kuhn: https://wiki.secondlife.com/wiki/Multi-Process_Client_VAG_--_draft
  • [13:52] Gareth Ellison: there's some code in the ArchWG SVN that's the start of a generic messaging system
  • [13:52] Tao Takashi: Zero: not really ;-)
  • [13:52] Rob Linden: [2] , I believe
  • [13:52] Saijanai Kuhn: I need to do that Gareth, Thanks for eminding
  • [13:52] Tao Takashi: and my old pysecondlife code is now at [3]
  • [13:52] Tao Takashi: hey, Google Code even has a wiki ;-)
  • [13:52] Gareth Ellison: [4]
  • [13:52] Saijanai Kuhn: *reminding
  • [13:53] Gareth Ellison: archwg
  • [13:53] Tao Takashi: Sai: help with what?
  • [13:53] Tao Takashi: Yak shaving? ;-)
  • [13:53] Saijanai Kuhn: setting up python script for this stuff
  • [13:54] Cleon Benelli: sup
  • [13:54] Tao Takashi: Rob: thanks!
  • [13:54] Saijanai Kuhn: not found [5]
  • [13:54] Gareth Ellison: let me double-check the link
  • [13:54] Tao Takashi: Sai: depends if I have time soon to actually start it ;-) But should I start I will tell you
  • [13:55] Tao Takashi: guess we will first discuss the protocol a bit more
  • [13:55] Tao Takashi: so are you going to release the cap server, too?
  • [13:55] Tao Takashi: and while we are on it, what about the server source code? :)
  • [13:55] Zero Linden: I think we should - Rob?
  • [13:56] Rob Linden: zero: don't let me stop you :)
  • [13:56] Zero Linden: The cap server is going to be a fundimental primitive, like cHTTP - so we should open up our implemetnation
  • [13:56] Zero Linden: Cool
  • [13:56] Gareth Ellison: grr, someone took my root access away on openmv.org and removed the viewvc installation....
  • [13:56] Tao Takashi: Sai: I will probably use the pysecondlife project for that stuff, too
  • [13:56] Tao Takashi: cool! :)
  • [13:56] Tree Kyomoon: Id really like to see how that works
  • [13:57] Tao Takashi: I think I would probably start without cHTTP and add it later where needed
  • [13:57] Tao Takashi: I am just not sure whether I go with eventlet/mulib or something in Zope ;-)
  • [13:57] Tao Takashi: or some mix
  • [13:58] Tao Takashi: but I also notice that I have too many projects again
  • [13:58] Zero Linden: Tao - don't we all!
  • [13:58] Tao Takashi: is your caps server in production right now?
  • [13:58] Saijanai Kuhn: has none, waiting on LInden HR to get back to him
  • [13:58] Saijanai Kuhn: the more ppython I can learn now, the better
  • [13:59] Tree Kyomoon: anyone have the first half of these office hours they can lay on me?
  • [13:59] Tao Takashi: you can also learn python without waiting for LL ;-)
  • [13:59] Gareth Ellison: i'm going to logout here and rant at whomever removed my root access on openmv.org - if anyone wants to point a standard SVN client the URL is svn://openmv.org/archwg
  • [13:59] Dzonatas Sol: quietly works on an overfull plate
  • [13:59] Saijanai Kuhn: exactly, Tai. The more pyton PROJECTS I can do now, the better
  • [13:59] Rob Linden: Tao: yeah, caps are in production.
  • [13:59] Zero Linden: Tao - "you're soaking in it right now"
  • [13:59] Tao Takashi: Rob: cool :)
  • [14:00] Tao Takashi: btw, when do we get Windlight? ;-)
  • [14:00] Rob Linden: (Zero: you're pop culture reference dates you ) :)
  • [14:00] Zero Linden: Tree - I have it all
  • [14:00] Saijanai Kuhn: if we could set up our own caps server to work with we could test things a lot faster
  • [14:00] Rob Linden: (and "regions" you)
  • [14:00] Zero Linden: I have a think for Madge
  • [14:00] Zero Linden: *thing
  • [14:00] Tree Kyomoon: if you want me to post it, madge, I can :)
  • [14:00] Tao Takashi: well, if we'd get that server we wouldn't have to code it ourselves, basically ;-)
  • [14:00] Gareth Ellison: is outta here - someone grab that lovely code i put up
  • [14:01] Tao Takashi: cya Gareth!
  • [14:01] Saijanai Kuhn: later gareth
  • [14:01] Zero Linden: well all
  • [14:01] Zero Linden: it is 2pm
  • [14:01] Tao Takashi: so now when is windlight coming? ;-)
  • [14:01] Zero Linden: and I have to run
  • [14:01] Saijanai Kuhn: Thanks Zero
  • [14:01] Rob Linden: me2
  • [14:01] Saijanai Kuhn: Thanks Rob
  • [14:01] Zero Linden: thanks all for coming
  • [14:01] Tao Takashi: once you ask the uneasy questions everybody needs to run ;-)
  • [14:01] Rex Cronon: bye zero
  • [14:01] Rob Linden: thanks all
  • [14:01] Morgaine Dinova: See you :-)
  • [14:01] Tao Takashi: thanks, Rob and Zero
  • [14:01] Rex Cronon: bye rob
  • [14:02] Finrod Meriman: thanks
  • [14:02] Tao Takashi: waiting for the caps server then ;-)
  • [14:02] Tara5 Oh: thanks Zero and Rob!
  • [14:02] Zero Linden: see you Thursday 7:30 this week - keep an eye pen for it to change next week
  • [14:02] Masaw Umaga: bye all
  • [14:02] Zero Linden: later