User:Zero Linden/Office Hours/2007 Oct 02

[13:02] Iridium Linden: Something free in Second Life? No way.
[13:02] Kooky Jetaime: crap, Rob and Periapse too
[13:02] Rob Linden: hi folks
[13:02] Morgaine Dinova: They are nice , yes. Would fit well into DarkLife theme
[13:02] otakup0pe Neumann: hey rob
[13:02] Tillie Ariantho: Heyas Rob. :)
[13:02] Morgaine Dinova: Hi Rob!
[13:02] McPhenius Swain: Hola Rob
[13:03] Tess Linden: position of a prim in a zone is held by the simulator
[13:03] otakup0pe Neumann: so uhh. how about those protocol changes yesterday.
[13:03] Tillie Ariantho: hums "upside down"
[13:03] Morgaine Dinova: Exclusively in the sim?
[13:03] Petronilla Paperdoll: has a column about freebies at the avastar, so please let her know about the free boots :-)
[13:03] Kooky Jetaime: Squirrel, good squirls don't wear dresses and sit like that
[13:03] otakup0pe Neumann: i don't think eddy will be able to sleep for days.
[13:03] Squirrel Wood: ^^
[13:03] Tess Linden: yes, it gets sent back to the client as object updates
[13:04] Eddy Stryker: yeah, are there some sort of "actual release notes" that get pushed out? the ones that would have made note of the message_template.msg changes, and the udp packets that were deprecated but not marked deprecated like AgentGroupDataUpdate
[13:04] Morgaine Dinova: And in what form is it held on the sim?
[13:04] Tess Linden: which protocol changes
[13:04] Eddy Stryker: tess: the ones i'm talking about
[13:04] Gigs Taggart: Yeah a server-side change log would be nice
[13:04] otakup0pe Neumann: AgentGroupDataUpdate comes in via caps event queue now.
[13:04] Kooky Jetaime: and why did I get an object return from an object that wasn't where it was when it was returned?
[13:05] Rex Cronon: hello everybody
[13:05] otakup0pe Neumann: which kinda broke libsecondlife.
[13:05] Morgaine Dinova: Hi Rex
[13:05] Rob Linden: Tess: Eddy sent mail to sldev with the diff (look for John Hurliman)
[13:05] Kooky Jetaime: Eddy has a point though, a nice Release Notes for new Server versions would be great, since we get them for the Client.
[13:06] Tess Linden: ah yes, which forwarded it to icehouse, i did see the change
[13:06] Tess Linden: Which*
[13:06] Eddy Stryker: message template changes can be dealt with, but things being silently deprecated is more difficult. i'm wondering if we are maybe dealing with things in the wrong way, are clients supposed to be able to transparently handle udp packets deprecating in favor of caps transports?
[13:06] Tess Linden: I guess we would be able to look at the capabilities list, of any client side new capabilities and match up to see what was *not* deprecated
[13:07] Tess Linden: I think it depends on why it was left in
[13:07] Gigs Taggart: so that it could be silently depricated!
[13:07] Tess Linden: I can imagine if we wanted to leave the old UDP code in just in case something goes wrong with the capabilities and deprecate it in a later release
[13:08] Gigs Taggart: it was announced that the plan was to migrate parts to caps without the necessity of client updates
[13:08] Eddy Stryker: ok, so for people writing their own clients (libsecondlife), we need to be able to handle situations like this transparently
[13:09] Tess Linden: that makes sense. if we deprecate the message, then older clients won't get those messages handled
[13:09] otakup0pe Neumann: Hehe yeah.
[13:09] McPhenius Swain: true
[13:09] Morgaine Dinova: Eddy, is libsl able to support that already, people using it as the front endpoint and adding visual or whatever as a backend?
[13:10] otakup0pe Neumann: If that's the direction we need to take that's cool. Prior notice would have been awesome though. fourtunately AvatarGroupsReplies still come in normally. hooray workarounds.
[13:10] otakup0pe Neumann: morgaine, yes
[13:10] Kooky Jetaime: Tess - simple fix is to run both caps & depreciated, then once you get a significant number of elements to take offline, push a required update
[13:10] otakup0pe Neumann: libsecondlife also poweres opensim
[13:10] Morgaine Dinova: That's superb, otaku
[13:10] Kooky Jetaime: Say once every 60 days max?
[13:10] Tess Linden: eddy: is the case you are talking about, we add a new capability, we dont deprecate the old code, and one day we deprecate it, and your libsecondlife client wont be able to support the new capabillity?
[13:10] Dizzy Banjo: A broader question : does LL see the opportunity of of grid expansion as a way of rethinking any aspects of the client / server relationship .. or is the work currently being done primarily focused on direct scaling of the existing SL experience to a different architecture ?
[13:10] otakup0pe Neumann: tess yeah that's what happened yesterday
[13:11] Tess Linden: why doesnt libsecondlife get the updated capabillity code?
[13:11] Eddy Stryker: tess: right now we have been hand-coding in support for each new capability as it is added to the grid
[13:12] Tess Linden: dizzy: what parts of the client/server relationship are you thinking of?
[13:12] Eddy Stryker: i think it is becoming apparent that this is not the correct way, and we should have a generic handler that can serialize llsd<->message_template format bi-directionally
[13:13] Eddy Stryker: so that any event received on the caps event queue can be serialized in to a format as described in message_template.msg, is that correct?
[13:13] Tess Linden: maybe if we had a web service that returned all the available client side capabillities by name?
[13:13] Eddy Stryker: that would be really great
[13:14] otakup0pe Neumann: that would be an awesome method for now, but it seems like long term we'll need something automagic like eddy said
[13:14] Dizzy Banjo: i dont know.. its an open ended question.. there are lots of things which i would personally like to see developed.. but my question is more whether that is currently part of this grid rethink discussion.. or is it just about directly scaling
[13:15] Gigs Taggart: Dizzy, the underlying assumption is that it has to be something we can get to incrementally. Other than that, it's open.
[13:15] Tess Linden: when we were designing message liberation, those hooks were discussed as an optional hook, but at the time we did not understand the use cases
[13:16] Tess Linden: the capabilities API may not necessarily be the same as the old message template for that message
[13:16] Eddy Stryker: just want to confirm, do caps that have the same name as packets in message_template.msg follow the same structure, or is it just coincidence that they share the same names?
[13:16] Eddy Stryker: argh. ok
[13:16] Tess Linden: a lot of the times, the developer who chose to move the capability over, chose that route because of the flexibillity LLSD has over the message template
[13:17] Gigs Taggart: Tess, Eddy, wouldn't a simple communication like a server-side changelog fix this? :)
[13:17] Eddy Stryker: so we need to hand-write new processing code for each capability, unlike the old udp system
[13:17] otakup0pe Neumann: yeah can we start with server side changelogs :|
[13:17] Saijanai Kuhn: in which case we need a LLSDMessageFactory method
[13:17] otakup0pe Neumann: grep out security stuff, that's understandable
[13:17] Saijanai Kuhn: class/whatever
[13:17] 57 Miles: hi everyone
[13:18] Veritas Variscan: Hi 57
[13:18] Rex Cronon: hi 57
[13:18] Deeeep Witte: Hi Nick
[13:18] Eddy Stryker: gigs: if i understood right, things are transitioning from a system where we have a file format detailing the structure of calls out, to a system where there is no documentation, no way of the client knowing what it is supposed to handle or not, and no idea what means what once you figure out the capabilities you are supposed to handle
[13:18] Tess Linden: There's so much that's checked into the code base, and a lot of merges happen between branches, if we log everything, it may get too cluttered and may not concern everyone
[13:18] Tess Linden: I do see the value in having good documentation linked to changes
[13:18] Gigs Taggart: Eddy, that's what I'm saying, primarily this seems like a communication/documentation problem to me
[13:19] Rob Linden: I'm not exactly sure where to insert this into the process, but there's probably some documentation we can put out on wiki.secondlife.com as things are getting checked in
[13:19] otakup0pe Neumann: i think eddy started protocol docs way back when. maybe just keeping those updated ?
[13:19] Deeeep Witte: A question on legislative life... what is the policy for custom worlds in SL now?
[13:20] Rob Linden: we just have to be careful not to confuse people with changes that are coming vs changes that are here
[13:20] Kooky Jetaime: Tess - Only you can't have someone come back and say "You didn't tell us".. its better to have to much than not enough, and better to have what you need instead of to much.
[13:20] Eddy Stryker: gigs: the udp protocol was self documenting (as best it could be). caps is not self documenting
[13:20] Deeeep Witte: how will legislative life link to the mian grid?
[13:20] Tess Linden: sounds like a callout for good API document for every capability that's added
[13:20] Gigs Taggart: What is "legislative life"?
[13:20] Eddy Stryker: so if it is not self documenting, at least manual documentation would be nice
[13:21] Tess Linden: the LLHTTPClient call used to invoke the capability does give a lot of information about parameters though
[13:21] Deeeep Witte: Newt Gingrich announced a virtual world for politicians!
[13:21] Rob Linden: I am concerned about a process that relies on developers being good about documentation, because...well...they aren't ;-)
[13:21] otakup0pe Neumann: Tru.
[13:21] Deeeep Witte: political virtual world: can you imagine the sex they will have ;-)
[13:21] Rob Linden: (that's a general industry wide truism)
[13:21] Zha Ewry: And nver will be. If we can get tolerable spec it will be a miracle
[13:21] Saijanai Kuhn: votes for well-documented messagefactory class
[13:22] Kooky Jetaime: Guess its a good thing I'm not in the industry.. I can document very well..heh
[13:22] Zha Ewry: I'll be happy, if every REST endpoint has a way of getting it's spec from it
[13:22] Deeeep Witte: seconds Saijanai's vote!
[13:22] Tillie Ariantho: the source code is a good documentation, too .)
[13:22] Saijanai Kuhn: LOLOLOL
[13:22] Rex Cronon: i agree sijanai
[13:22] Rex Cronon: saijanai*
[13:22] Eddy Stryker: zha: that would be very nice
[13:22] Gigs Taggart: that's not a bad idea.. could parts of the server source be published in relation to this?
[13:22] Saijanai Kuhn: immitates three stooges on floor going in circles and whooping (note to self, create animation for this)
[13:22] Gigs Taggart: just the communications/caps bits? Is it separate enough?
[13:23] otakup0pe Neumann: tess you said something about a discovery web service a little while back
[13:23] Tess Linden: its partr of the client isnt it?
[13:23] Zha Ewry: REST gives us sane ways of marking off discovery, it you want to
[13:23] Gigs Taggart: Zha, sane ways are boring
[13:23] Tess Linden: the new clients released invokes these new capabilities with a very clear API
[13:23] Kooky Jetaime: this is a bit to techy for me today and I have other things that I need to do, so I'll see you all later
[13:24] Baba Yamamoto: aw crap i'm late
[13:24] Rex Cronon: what .cpp class has this API?
[13:24] Iridium Linden: Bye, Kooky.
[13:24] Zha Ewry: Which API? Zero's been pushing the notion of definign the parts at the Web Service level
[13:25] Rex Cronon: where is specification doc for this API?
[13:25] Eddy Stryker: Tess: if the new CAPS stuff doesn't necessarily follow message_template.msg (even though it does. so far...) what about publishing a caps_template.msg format either as a file or through a web service so clients can download enough information to know how to represent the CAPS calls in code
[13:25] Tess Linden: rex: whichever one is invoking the capablilities, you can search for getCapability(), each call that invokes those capabilities do so with certain arguments
[13:25] otakup0pe Neumann: coughs
[13:26] Gigs Taggart: heh, caps_template.msg :)
[13:27] Squirrel Wood: /ao on
[13:27] Tess Linden: like rob said, it is hard for developers to get into the habbit of documentation, but I think the capabillity calls within the client are very selfdocumenting
[13:27] Baba Yamamoto: how about not caps_template.msg ;0 we just need unencumbered documentation of some kind
[13:27] Tess Linden: the httpclient::host/get/put calls put an llsd witha specific template
[13:27] Tess Linden: and the response class parses the llsd response
[13:27] Morgaine Dinova: It isn't hard for devs to get into the habit of documenting. They just need a mild incentive ... like a major bollocking. :-)
[13:28] Gigs Taggart: flogging will continue until morale improves
[13:28] Morgaine Dinova: Hehe
[13:28] Zha Ewry: Well, it won't be a public spec until people can read more than GPL liscnesed code to find out how to use it
[13:28] Rex Cronon: i think java is a good example how documentation can be included in the source code, and how from source code html help files can be genereated
[13:29] Tess Linden: I know, if I'm tasked with documenting the API of a specific capability,I would be looking at the httpclient call, and the responder to determine what the API to put into the wiki
[13:29] Squirrel Wood: wonky AO script :(
[13:29] Eddy Stryker: if the documentation for all of this is GPL licensed C++ source code we will probably setup an official clean room reverse engineering team
[13:29] Rob Linden: saying that it's "not public" because it's "only" in GPL source code isnt fair
[13:29] Rob Linden: it's public
[13:29] Zha Ewry: Yeah it is
[13:29] Morgaine Dinova: Documenting what you are doing, in an appropriate matter, which can be incredibly terse but informative, is part of working professionally.
[13:30] Zha Ewry: If you put it in a form that limits who you can us it, it isn't public
[13:30] Gigs Taggart: You don't have to cleanroom GPL.... FFS just don't rip off code and you'll be fine.
[13:30] Gigs Taggart: It's not really a technical problem that you can't trust your own coders not to steal code.
[13:30] Tess Linden: hm, k i dont really understand the GPL not being public part
[13:30] Baba Yamamoto: Rob, we can't deal with GPL
[13:30] Saijanai Kuhn: in the case of the client, please do NOT rip off the code. That would be futile
[13:31] Zha Ewry: Tess, the GPL explicitly messes with how prople csn use the code
[13:31] Zha Ewry: That's not public
[13:31] Rob Linden: ummmm....it's public Zha
[13:31] Veritas Variscan: /must run, take care everyone!
[13:31] Baba Yamamoto: it's public.. it's just encumbered ..
[13:31] Zha Ewry: Encumbered ot th epoint where if a debve;loper reads it
[13:31] Zha Ewry: they can be sued is not public
[13:31] Saijanai Kuhn: its not private, but its not copyable for commercial use
[13:31] Eddy Stryker: Rob: so we can look through the SL viewer source code to learn how things work and reimplement it in BSD licensed libsecondlife correct?
[13:31] Squirrel Wood: whoever is wearing that "::: CD ::: Bishounen AO (Wear Me)" please take it off.
[13:31] Rob Linden: Eddy, as long as you don't steal code, yes
[13:32] Eddy Stryker: thank you
[13:32] otakup0pe Neumann: can we get that in writing ?
[13:32] Eddy Stryker: we can't steal code since the lib is in C#, so that should not be a problem
[13:32] Baba Yamamoto: Can we get a lawyerspeak license exception?
[13:32] Baba Yamamoto: ;)
[13:32] Rob Linden: no
[13:32] otakup0pe Neumann: hooray cleanroom !
[13:32] Morgaine Dinova: The GPL does not mess with how people can use code, please don't spread myths. It is not a usage license. The GPL is a distribution license, and nothing else.
[13:32] Rob Linden: there's been a ton written about the GPL
[13:32] Baba Yamamoto: Rob, a ton.. and everyone has an opinion
[13:32] Baba Yamamoto: it's all different
[13:32] Rob Linden: we're not going to get into the business of reinterpreting it for people
[13:33] Tess Linden: hm, so this whole discussion about documenting the capability API is just so people can steal code? am i missing something?
[13:33] otakup0pe Neumann: it seems to have devolved to that.
[13:33] Rex Cronon: u don't need documentation to steal code
[13:33] Zha Ewry: No.In fact, declaining "It is ok, we've dcoumented the template" is pretty much a co out
[13:33] Eddy Stryker: ok, as long as we're not going to get sued because of "how did you learn how that works!"
[13:33] Morgaine Dinova: Correct, don't interpret it yourselves, just listen to Eben Moglen's words, the barrister who wrote the GPL: "The GPL is NOT A USAGE LICENSE".
[13:33] Rob Linden: there are some paranoid lawyers that advise not reading GPL code for fear of "contamination". Yet, I wonder what those same lawyers would say about reading....oh, say Microsoft sample code
[13:33] Rex Cronon: if u want to create something new u do need documentation
[13:33] Gigs Taggart: :)
[13:34] Saijanai Kuhn: always thought that that was the oroginal purpose of GPL: to provide an example of 'How things work"
[13:34] Squirrel Wood: There are only so many ways one can code a "hello world" program efficiently. In the end, the code will end up similar to someone elses code. if not identical.
[13:34] otakup0pe Neumann: i'm still in favour of a web service that publishes info on caps.
[13:34] Tillie Ariantho: I thought SL is going to be an open platform. Now you worry about code being stolen? If you really fear THAT, then just do NOT release code but a well documented API...
[13:34] Morgaine Dinova: Tillie: agree
[13:34] Saijanai Kuhn: I can't imagine anyone from LL worrying about the client code being stolen...
[13:34] Tess Linden: I think you guys are the ones who brought up GPL in the first place
[13:35] Gigs Taggart: Tillie these peole and companies themselves are worried about infringing.
[13:35] Rob Linden: Tillie: we provide the source code under a very liberal license, but we still reserve some rights
[13:35] Tillie Ariantho: You want people to contribute into the SL code, but if people want to have a look in and use some of it, it is now nice again... that's pretty weird.
[13:35] Rob Linden: Tillie: you can use it, under the terms of the GPL
[13:35] Tillie Ariantho: now -> not
[13:35] Gigs Taggart: they can use code verbatim, if they comply with the GPL, otherwise it's look but don't copy :)
[13:36] Rob Linden: what this discussion is about is that there are people here who don't want to comply with the GPL
[13:36] Morgaine Dinova: Yep. Or even better, phone up Eben Moglen.
[13:36] Gigs Taggart: Rob and I think that's fine that they don't want to use the GPL.
[13:36] Rob Linden: ...which is fine, except they can't copy our code
[13:36] Tillie Ariantho: ok
[13:36] Saijanai Kuhn: wow, that's the exact opposite of how I thought the GPL was suposed to work: use as-is, or rob ideas, but don't do both
[13:36] Morgaine Dinova: That's a clear statement. Thank you Rob.
[13:37] Rex Cronon: i have no interest incopying anoybodys code
[13:37] Zha Ewry: But.. then.. you do have a problem, becase for better or worse, and I am not a lawyer.. but.. I lhave to listen to coprorate ones, people do orutinely say "Do not look at GPL code for fear of contamintaion"
[13:37] Zha Ewry: You can say "go talk to Eben"
[13:37] Gigs Taggart: Get better lawyers :)
[13:37] otakup0pe Neumann: This was a concern within the openim project as well.
[13:37] Rob Linden: what Gigs said :)
[13:37] otakup0pe Neumann: heh
[13:37] Zha Ewry: And I like Eben.. but lot sof coproate players are going to feel you are telling them "don't play"
[13:38] Eddy Stryker: i don't think anyone here has any desire to copy source code out of the second life viewer. i am just apprehensive about the fact that the only documentation for the protocol now exists in the viewer source code
[13:38] otakup0pe Neumann: yeah...
[13:38] Gigs Taggart: Zha if libsl "documents" this stuff in BSD form in a way that is pretty hard to infringe, since it's a different programming language, then you are set.
[13:38] Zha Ewry: And.. In ANY CASE.. The reality is that reading source to understand a public spec, is just not on
[13:38] otakup0pe Neumann: we just need more interns to document it to the wiki.
[13:38] Morgaine Dinova: I can understand IBM paranoia, after the SCO silliness. But you can get *too* paranoid :-)))
[13:38] Deeeep Witte: clearer documentation would make it easier for smaller developers to write small apps
[13:39] Rob Linden: so.....anyone here should feel free to write up documentation based on the source code and put it on the wiki
[13:39] Tillie Ariantho: Like Prokofy. .P
[13:39] Zha Ewry: Actually, I can cope with my own laeyers
[13:39] Sifu Moraga: You can't get too paranoid if you are as big as IBM
[13:39] Eddy Stryker: i've started the documentation effort at [1] (still empty so far)
[13:39] Zha Ewry: But.. GPL for better or worse stirs up corporate lawyers
[13:39] Tess Linden: can we give love to os devs?
[13:39] Deeeep Witte: looks around " Who said paranoid?"
[13:39] Rob Linden: Eddy: we may want to merge that with another page (/me looks)
[13:39] Baba Yamamoto: heh BSD is so worry free..
[13:40] otakup0pe Neumann: Yeh there is some information elsewhere I think.
[13:40] otakup0pe Neumann: [2]
[13:40] Rob Linden: that's the one
[13:40] Rex Cronon: would be very helpfull if there was a UML diagram, for the viewer.
[13:40] Zha Ewry: Rex, it would make your eyes bleed
[13:40] Eddy Stryker: we also need to hash out a standard way of documenting CAPS structures. llsd notation format might be close enough
[13:41] otakup0pe Neumann: good enough for a start, at least
[13:42] Eddy Stryker: rob: i started a new page because that one got in to a lot of things like file formats that are dealt with on the client side and have nothing to do with the protocol. it seemed to be more geared towards documenting the sl viewer specifically
[13:42] Gigs Taggart: UML for the viewer would be pretty scary stuff
[13:42] Rex Cronon: but the code is not good either to use as documentation:(
[13:43] Deeeep Witte: thinks of 3d SL UML....
[13:43] Rob Linden: Eddy: [3]
[13:43] Rob Linden: go ahead and hijack Protocol
[13:43] Iridium Linden: lol, Rob.
[13:44] Saijanai Kuhn: Dale Glass has the entire client code Doxygenated on his website
[13:44] Eddy Stryker: ok
[13:45] Rob Linden: is there a protocol documentation team we could form out of a subset of this group?
[13:45] Rob Linden: doesn't want this to just be on Eddy
[13:45] otakup0pe Neumann: needs more interns
[13:46] otakup0pe Neumann: i'll help eddy out
[13:46] Zha Ewry: would be glad to dive in on that
[13:46] otakup0pe Neumann: we need someone from ll
[13:46] otakup0pe Neumann: or it would make sense anyway
[13:46] Saijanai Kuhn: is still coding his silly floater. Will be availabe after that
[13:46] Tillie Ariantho: [4] <-- there it is
[13:46] Rob Linden: I'll push for more LL involvement, but remember that LL isn't the one asking for this.
[13:47] Saijanai Kuhn: LL has no need for better docs?
[13:47] Eddy Stryker: understandable, they already have internal documentation on all of this. or lindens "just know" how it works
[13:47] Sifu Moraga: wonders if I could get some students interested
[13:47] otakup0pe Neumann: nope. but i think ll is one of the groups asking for a grid architecture working group.
[13:47] Tess Linden: er, we look at the code...
[13:47] Rob Linden: we're too busy changing it to actually document it ;-)
[13:47] otakup0pe Neumann: heh
[13:47] Tess Linden: :P
[13:47] otakup0pe Neumann: i'd love to see the internal wiki.
[13:47] Tillie Ariantho: But Rob, we all want the Second Life platform to be a success, right? LL too...? So we all have to work together. And docs is one thing.
[13:47] otakup0pe Neumann: <_<
[13:47] Peter Newell: that's such a wise software development path! :p
[13:47] Baba Yamamoto: make donovan do it!
[13:47] Baba Yamamoto: ;)
[13:48] Saijanai Kuhn: first code, it, then document what you meant to do, eh?
[13:48] Tess Linden: it is easier though, having had worked on some of this stuff, and having more of the backend knowledge
[13:48] Eddy Stryker: hey peter :)
[13:48] Baba Yamamoto: ;0
[13:48] Rob Linden: seriously, though, I'll push for LL help on this
[13:48] otakup0pe Neumann: thanks rob
[13:48] otakup0pe Neumann: haha <3 donovan
[13:48] Saijanai Kuhn: recalls parsing through the 80+ functions of the llScrollListCtrl class...
[13:48] Zha Ewry: Really appreciated Rob, and we do konw what it's like to ask developers to do documentation
[13:48] Rob Linden: ....and I'm always in there futzing, so I'm a de facto volunteer
[13:48] Tillie Ariantho: hehe
[13:49] Eddy Stryker: i'll merge those wiki pages together and send a note out to sldev
[13:49] Zha Ewry: recalls a long deiscussion invoving nested for loops, and set of 12 single letter varibales. and the lack of comments
[13:49] Tess Linden: it would scale better if documentation came along with impelmentation, just have to get people into the habit
[13:49] Rob Linden: cool, thanks Eddy
[13:49] Baba Yamamoto: heh
[13:49] Saijanai Kuhn: sometimes, documentation helps BEFORE implementation...
[13:49] Tess Linden: we talked about including test plans with every checkin in Icehouse
[13:50] Tess Linden: and thats just one type of documentation
[13:50] Baba Yamamoto: should we start a documentation portal? ;)
[13:50] Tess Linden: theres also design documentation
[13:50] Gigs Taggart: Zha, what's the problem, that still leave 14 more possible variables to use. :)
[13:50] Tess Linden: which should include protocols too
[13:50] Rob Linden: I think the more that we can show that there's benefit to using the external wiki to the internal wiki, the more people at LL will use it
[13:50] Tess Linden: yes!
[13:50] Zha Ewry: If test plans are well worked out, they are deifnitely a good way of capturing intent
[13:50] Rob Linden: one big benefit is if there's a bunch of people externally cleaning up and augmenting
[13:50] Tess Linden: I just built my first linux client last week
[13:50] otakup0pe Neumann: that's a good way of looking at it rob
[13:50] Baba Yamamoto: rob, you know hwo to make them use the external wiki?
[13:50] Morgaine Dinova: Well every time you hire someone, the standing instruction should be "Document anything you learn" ... and if they have learned nothing then they do't get paid :-)
[13:50] Baba Yamamoto: ;) turn off the internal one!
[13:50] Iridium Linden: There's an external wiki?
[13:51] Tess Linden: and the internal wiki page said "look at the external wiki"
[13:51] Rob Linden: lol
[13:51] Iridium Linden: Just kidding.
[13:51] Eddy Stryker: if we want to start a whole new tangent, does LL have any sort of test suite for LLSD? there are three serialization formats and very nice docs on the wiki, but a comprehensive unit test of snippets of llsd would be a great project for someone if it doesn't exist yet
[13:51] Saijanai Kuhn: what's a wiki?
[13:51] otakup0pe Neumann: heh.
[13:51] Baba Yamamoto: ;0
[13:51] Saijanai Kuhn: <3 LLSD<=> xml
[13:51] Tess Linden: lemme tell you, i started following the steps to that without realizing i could just build the internal client, and its inspiring me to build the external one.... when I have.. time...
[13:52] Zha Ewry: Quasi serisously, tho, if you expect to end up with a spec that can be used, longer term, you will, in fact, need every blasted service documented, and not in code.
[13:52] Morgaine Dinova: There's a new Linden just dedicated to testing, we had an interesting Office Hours about it. It revealed that there is currently no auto or regression testing framework.
[13:52] Zha Ewry: And to a level that isn't "This is how LL coded it"
[13:52] Tess Linden: the fact that the internal wiki pointed to the external one is fantastic rob, I think we should secretly start doing that more to get people to use it more
[13:53] Deeeep Witte: what is llsd?
[13:53] Rob Linden: Tess: heh...."secretly" .....hehe
[13:53] Baba Yamamoto: we need LLSD -> JSON transform! ;0 is that too much of a tangent?
[13:53] Saijanai Kuhn: recalls parsing the 80+ functions of the llScrollListCtrl class... Docuemtation? We don't need no sinking documentation. Everyone knows that if you set the id in the LLSD, this overrides the value
[13:53] Sifu Moraga: Actually, I like the idea of unit tests as documentation. Usually works for me better
[13:53] Tess Linden: I have people ask me questions about capabilities all the time
[13:53] Zha Ewry: >sifu: as long as the unit tests capture the full design, not the as coded desing
[13:54] Sifu Moraga: true, Zha
[13:54] otakup0pe Neumann: baba didin't i see donovan tweet about that ?
[13:54] Eddy Stryker: baba: the notation format of llsd is so close to json there's no point, the only thing you get converting over is losing strongly typed data like uuids
[13:54] Tess Linden: rob: if we put up titles on the internal wiki (that developers search against) then point the internal pages to the external pages, thats more secret right?
[13:54] Sifu Moraga: That is usually the problem with unit tests as documentation
[13:54] Baba Yamamoto: you can always start behavior testing ;0 !!
[13:54] otakup0pe Neumann: he said something about json
[13:54] Baba Yamamoto: otakup0pe, i dunno
[13:55] Zha Ewry: The fear one has.. in evolvig this out of working code, is endeding up not with an architecture, but rather a codification of a speicific instance of that architetcure
[13:55] Rob Linden: Tess: as it turns out, our internal search engine actually crawls the external wiki already. but you're right about more outbound links
[13:55] Zha Ewry: (Which may sound like nit picking but really isn't)
[13:55] Saijanai Kuhn: nods. Like GUIs that update based on teh current framerate of the viewer
[13:55] Tess Linden: oh, i didnt know that, cool
[13:55] Eddy Stryker: zha: true, but until there are real working alternative implementations that's all we'll likely have
[13:55] Tree Kyomoon: can someone pass me the IM history?
[13:56] Rob Linden: Tree: sure
[13:56] Iridium Linden: is off to a meeting. *waves*
[13:56] Zha Ewry: pauses to count lindens, and is impressed
[13:57] Tree Kyomoon: hey Iridium!
[13:57] Saijanai Kuhn: Eddy: but until there is a spec divorced from the implementation, you can't have an laternate implementation
[13:57] Tess Linden: how many lindens does it take to make up for zero linden hehe
[13:57] otakup0pe Neumann: hehe
[13:57] Morgaine Dinova: LOL
[13:57] Zha Ewry: LOL
[13:57] Rex Cronon: bye iridium
[13:57] Zha Ewry: Many.
[13:57] In Kenzo: waves and thanks the Ls for trying to = zero
[13:57] Morgaine Dinova: That's a great line, hehe
[13:57] Squirrel Wood: :)
[13:57] Morgaine Dinova: Bye Iri
[13:57] Tree Kyomoon: multipy any linden by zero and you have it!
[13:57] Zha Ewry: Of courrse this tribe counts in Binary, Zero, One, Many.
[13:57] Saijanai Kuhn: Zero is the basic building block of set theory...
[13:57] Baba Yamamoto: hah with libsecondlife and opensim the grid has an open relationship.. ;0
[13:58] In Kenzo: polydigitery?
[13:58] Saijanai Kuhn: NULL, {NULL}, {{NULL}, NULL}...
[13:58] Rex Cronon: i think there is a group called "divided by one, united by zero" :)
[13:59] Rex Cronon: or is it the other way around?
[13:59] Zha Ewry: Also.. anyone who's chasing AWG stuff is welcome to bang on me for an invite to the AWGroupies.. group for tech talk about all of htis
[13:59] otakup0pe Neumann: ok dinner time on the east coast.
[13:59] Eddy Stryker: zha: k
[13:59] otakup0pe Neumann: i'm out. unlessthere are any other messages from ll :o
[13:59] Tess Linden: well, thank you everyone for coming by
[13:59] Zha Ewry: Early dinner, Otak...
[13:59] otakup0pe Neumann: can i haz windlight ?
[13:59] otakup0pe Neumann: zha it's 1800 :o
[14:00] Baba Yamamoto: EST + 1
[14:00] Zha Ewry: Ah
[14:00] Tess Linden: see y'all later!
[14:00] Periapse Linden: Thanks, Tess, for moderating in Zero's stead
[14:00] Zha Ewry: That's the *very( east coast
[14:00] Periapse Linden: Bye everybody!
[14:00] Rex Cronon: bye tess
[14:00] Sifu Moraga: cu!
[14:00] Zha Ewry: And.. Many thinks for all the Lindens
[14:00] Dizzy Banjo: thanks everyone, sorry i couldnt contribute more to this discussion other than my conceptual ramblings
[14:00] otakup0pe Neumann: isnt that the ISO standard dinner time ?
[14:00] Eddy Stryker: thanks tess
[14:00] Saijanai Kuhn: laters
[14:00] Turtle: Yeah,: bye...
[14:00] Tess Linden: :) bye!

Made with Tree's Free Wikifier

User:Zero Linden/Office Hours/2007 Oct 02

Navigation menu

Search