[9:28] Tao Takashi: so basically I have full concentration for 30 mins, after that my alt needs to attend another event and I might be more active there
[9:43] Zha Ewry: And. I have two small topics, if we run out of things to pester you and Tao about
[9:43] Saijanai Kuhn: watches the Happy LInden have a panic attack
[9:44] Enus Linden: so we are considering the merits of using zca in the pyogp package
[9:44] Movies1963 Beck: I've heard a rumor I'd like to see cleared up
[9:44] Enus Linden: weighing the flexibility it allows vs the use of for lack of a more reasonable term 'stock' python classes etc
[9:45] Enus Linden: pyogp aims to be 1. a client library for interacting with Sl and SL like agent and region domains 2. a test harness for testing itself and an implementation of an SL/opensim like grid
[9:45] Enus Linden: i'd also like to have in the package the ability to test agent and region domain implementations as well
[9:46] Enus Linden: ZCA as i understand it adds flexibility when incorporated, by defining apis for modules
[9:46] Enus Linden: each module can be swapped out for other modules that adhere to the api
[9:48] Tao Takashi: but it should make refactoring easier as it's hopefully not too big components we handle then
[9:48] Enus Linden: the only disavantage I see is adding a layer or abstraction in interacting with the classes, but this is a win as Tao points out in that we have to formally define things as we go
[9:52] Tao Takashi: I admit I also needed to get used to it first as it's the same with eggs and buildout (two things I also would like to use and use already in the Google code stuff)
[9:52] Enus Linden: Locklainn and I chatted earlier,he's happy to go with whatever we collectively decide
[9:52] Enus Linden: Tao, why would you like to see eggs and buildout?
[9:52] Juntje Allen: class Person(object): """a person""" def __init__(self, firstname, lastname, address, zip, city, dob): """initialize the address""" self.firstname = firstname self.lastname = lastname self.address = address self.zip = zip self.city = city self.dob = dob def fullname(self): return "%s %s" %(self.firstname, self.lastname)
[9:52] Tao Takashi: for instance with buildout you might not need that setup_path stuff I saw in your code
[9:52] Juntje Allen: is a sign up script i suppose?
[9:53] Tao Takashi: because buildout lets you automatically define what get's installed in a development sandbox
[9:53] Saijanai Kuhn: XML or script, I believe, Juntje
[9:53] Tao Takashi: so you can install certain versions of packages local to your development sandbox and don't need to pollute your global python installation with it
[9:54] Enus Linden: right, i perused that over coffee this am lol
[9:54] Tao Takashi: Juntje: Well, maybe, if they run in different sandboxes on a machine. But it's really more about different projects you develop on one machine
[9:55] Tao Takashi: so with buildout you have sandboxes with different versions if you want to
[9:55] Enus Linden: so i'm leaning toward voting yes on zca, despite the odd layer of abstraction it provides. i do think what gets built would require awesome and clear documentation and samples were we to go that route...
[9:55] Tao Takashi: buildout is sort of like make, maybe on a different level
[9:55] LaPiscean Liberty: Is that like Tribals build on you own machine and publis to grid
[9:56] Saijanai Kuhn: I'm all for it. I don't see that it makes the system any less language independent since we're not doing anything that other OOP languages don't support in their own way
[9:57] Tao Takashi: well, the library is language dependant anyway.. and ZCA does not interfere with the protocol itself
[9:57] Tao Takashi: it's just a question of how it's implemented
[9:57] Saijanai Kuhn: and we can add C or C++ modules for testing, and the itnernal implementations can still be straightfoward implementations of the protocols
[9:57] Enus Linden: but a python enabled (other language) module could be more easily plugged in as needed by a dev
[9:57] Tao Takashi: I also assume though that it needs some work to find the right structure. I played around with things a bit but that is very probably not the end result, many questions on how to model things are still open
[10:03] Tao Takashi: and I think we need some sort of dummy agent domain which responds to a login with maybe static data
[10:03] Tao Takashi: and returns some caps which also can be quite static
[10:03] Tao Takashi: and maybe some error conditions we can test against
[10:03] Enus Linden: so, say we do zca, buildout and eggs. we need 1. to overwrite the svn repo with a dir structure 2. publish docs explaining what's there 3. plan on next steps
[10:04] Tao Takashi: I am still not sure about the contrib agreement though, it makes contributions harder
[10:04] Enus Linden: i can't avoid that right now at all Tao.
[10:04] Enus Linden: is that a barrier for you? Sai, you?
[10:04] Enus Linden: Gareth is the only other person I know of who actively wants to participate with potentail contribs
[10:05] Tao Takashi: as I can always fork I guess it's not but the less control by LL the better IMHO
[10:05] Enus Linden: Locklainn and Infinity, plus Tess and Leyla, would contribute internally
[10:05] Tao Takashi: we also might need to setup some coding procedures, like how to code and what you shouldn't do etc.
[10:05] Tao Takashi: like write tests (is what you _should_ do ;-) )
[10:05] Enus Linden: yes Tao, and patterns + coding 'standards' loosely defined would be nice, so we're all speaking the same language
[10:16] Enus Linden: i'll start on 'spec' and roadmap and will make links available for discussion
[10:16] Enus Linden: i'd like to have a separate office hour that i run available for this and other topics... is friday at 10 sl time reasonable for those interested?
[10:18] Enus Linden: ah yes Tao. I need to talk to robla on how to check about contributor agreement signings, and can enable commit access for individuals
[10:18] Zha Ewry: Wow, Layla, are you wearing those booots, or are they wearin you?
[10:19] Enus Linden: I'll take care of that today or tomorrow
[10:19] Leyla Linden: hah, possibly the latter, they're from way back when i first started here
[10:25] Tao Takashi: well, for now I think we mainly concentrate on the test harness which can test whatever external agent or region domain (to what extend it's specified in whatever protocol that is)
[10:26] Tao Takashi: and we might need some library for doing these tests which is then pyogp.lib.*
[10:26] Tao Takashi: well, I was thinking about it and started some code but not much yet.
[10:27] Zha Ewry: The other. which is a more interesting question.. is a polite request, to start thinkign about how we want to manage a "safe" test grid in terms of content
[10:28] Zha Ewry: When we're doing interop testing, once we'll move stuff?
[10:29] Zha Ewry: I'll be using an Alt, with no content from any SL stores, just default systemn stuff
[10:29] Zha Ewry: But.. ideally, we'd have a small grid, with an asset server, which only has that stuff on it
[10:29] Zha Ewry: So, it doesn't even become an issue
[10:30] Zha Ewry: Since, we're goign to sort out all those issues in time, it would be nice to not feed the fears of the content creators. nor get sucked into a sideshow aboutr that
[10:31] Saijanai Kuhn: is a great barker for sideshows...
[10:31] Zha Ewry: Yes, you've held a lovely one or two in Jira, this month
[10:31] Saijanai Kuhn: amazing noise to signl ratio there
[10:32] Tao Takashi: well, do we need some contribution agreement for such content?
[10:32] BlueWall Slade: word is that Fashion Research Institute will be providing some content for OSGrid....
[10:32] Zha Ewry: I think the goal, here, is just to get the thought on the table, and keep it somewhat in mind
[10:32] Tao Takashi: we can also make some call for content
[10:32] BlueWall Slade: so you might even find some willing content developers to participate
[10:32] Tao Takashi: or I can simply provide my stuff ;-)
[10:33] Tao Takashi: well, I think it won't be such a big problem but you are right, we should keep this in mind and not suddenly copy whatever stuff from SL to this grid
[10:33] Zha Ewry: I don't think it is a serious problem, if we're at all careful
[10:34] Zha Ewry: What I'd like to avoid, is the sudden posting of "Look at me on OpenSim in my Linden SL content" snaps which would be contentious
[10:38] Zha Ewry: You can have plenty of openSim locla content
[10:38] Zha Ewry: what we don't want to do, is , int he course, of working out and testign asset stuff, to end up copying stuff which isn't intended to move
[10:39] Rex Cronon: is there any way right now, for somebody to copy scripts from here to other grids? my understanding is that u can't do that yet. right?
[10:39] Zha Ewry: I'm basically going to be reminding people of this, over the next few weeks, to get it into everyone's head al ittle bit
[10:39] Zha Ewry: And, no, there isn't Rex, tho, at some point, that becomes an interestign question
[10:39] Enus Linden: has to run, but would love to see where the content discussion goes. will someone post a transcript to wiki please?
[10:40] Tao Takashi: take care Enus, thanks for managing the pyogp stuff :)
[10:40] Zha Ewry: I think we'r emostly done, for now
[10:40] Zha Ewry: Just raising people's awareness ;-)