User:Which Linden/Office Hours/2008 Mar 5

From Second Life Wiki
< User:Which Linden/Office Hours
Revision as of 12:28, 3 July 2008 by Which Linden (talk | contribs) (New page: * [11:11] Which Linden: hmm....aw groupies chat dead * [11:11] Which Linden: that is le sad * [11:11] Which Linden: d8d...)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search
  • [11:11] Which Linden: hmm....aw groupies chat dead
  • [11:11] Which Linden: that is le sad
  • [11:11] Which Linden: d8dead84-8180-edce-bbe8-e04b7600d35c
  • [11:13] Saijanai Kuhn: hey guy
  • [11:13] Which Linden: hey Sai
  • [11:13] Which Linden: did you know, we found and fixed the bug with the group IMs being delivered with the wrong name?
  • [11:13] Saijanai Kuhn: some guy on libsl-dev bragged he'd gotten a python chat bot working
  • [11:14] Saijanai Kuhn: yeah, I saw that. Suddenly my bot stopped working at all ;-)
  • [11:14] Which Linden: hah
  • [11:14] Which Linden: um, I really apologize for not looking at your code
  • [11:14] Saijanai Kuhn: the bot ONLY worked when the bug popped up
  • [11:14] Which Linden: I need a few hours to do it justice, and I haven't had a few hours free
  • [11:14] Which Linden: ...and now group chat to AW groupies isn't working for me
  • [11:15] Which Linden: is it working for you?
  • [11:15] Saijanai Kuhn: sure I understand. Though,mk I think that it might have had a clue for where the bug was since it was a 100% repro of that bug
  • [11:15] Saijanai Kuhn: just only once per session, however
  • [11:16] Saijanai Kuhn: anyway, this guy claimed he had worked up gorup IM purely in Python and when I begged to let me see the code he sputtered and logged off. O well
  • [11:16] Which Linden: that is not very kindly of him
  • [11:17] Saijanai Kuhn: I think he was not being very truthful, to be honest. Claimed he'd gotten in working in 10 minutes or something. POSSIBLE, in theoyr, but not likely
  • [11:17] Which Linden: It'll be nice when we have a nice clean restful api for group im
  • [11:18] Which Linden: Did you see that Donovan checked in the capability server?
  • [11:18] Saijanai Kuhn: I saw the work you guys did on the strawman API docs. Very nice. It kinda is ironic since I'm trying for a tech writer job right now and the current login page was going to be my demo of my l33t skills
  • [11:18] Saijanai Kuhn: No I didn't
  • [11:18] Which Linden: [1]
  • [11:19] Which Linden: Yeah, the login api thing is interesting cause it has to cover both the current implementation, warts and all, and where we'd like to take it
  • [11:19] Saijanai Kuhn: that's not hte long-poll server for EventQUeueGet though, I think
  • [11:19] Which Linden: No, that server is in the simulator
  • [11:19] Saijanai Kuhn: I'm trying to tidy up a few missing "{" and whatnot and extend it and take it back to the current login page
  • [11:20] Saijanai Kuhn: its not quite complete enough to write a script or program from scratch right now
  • [11:20] Saijanai Kuhn: which was the critereon ZHa and I set up for the AWG protocol docs
  • [11:20] Which Linden: What's needed to make it complete?
  • [11:20] Saijanai Kuhn: I write it in Python, document it, and she tries to take my docs and rewrite it in Java as a test of completeness and language-neutrality
  • [11:21] Which Linden: That's a great collaboration there
  • [11:21] Saijanai Kuhn: the NEW part looks to be more complete. You guys just gave an abreviated version for the current stuff though
  • [11:21] Which Linden: Heh
  • [11:22] Which Linden: Kinda indicative of how we feel about it, I suppose
  • [11:22] Saijanai Kuhn: so, I can still show off my writing siills with something I'm familiar with by redoing the current protocol using your template
  • [11:22] Saijanai Kuhn: there's a few missing braces and so on in the new sectio/legacy section though
  • [11:23] Which Linden: For minor errors like that you should just edit
  • [11:24] Saijanai Kuhn: right. A couple of places, I'm not quite sure what you're trying to say though
  • [11:24] Which Linden: Heh, it's not me who wrote it, so I can't be sure either
  • [11:24] Which Linden: Blame Tess. :-)
  • [11:24] Saijanai Kuhn: https://wiki.secondlife.com/wiki/Second_Life_Login_API_Strawman#Response like under the "reason" response
  • [11:25] Saijanai Kuhn: I've got her in IRC but she seems to be afk right now
  • [11:26] Which Linden: OK, so that section seems pretty clear to me, except that it's kinda ambiguous what you're supposed to do if you get the "multiple accounts" failure
  • [11:27] Which Linden: Oh, I see, there's that "other reasons" section that doesn't really explain how those reasons will appear in the response
  • [11:27] Saijanai Kuhn: well, the dictionary/map entry seems to be missing a delimiter.
  • [11:27] Morgaine Dinova: Greetings :-)
  • [11:28] Saijanai Kuhn: Hey MOrgaine. We're nitpicking the new strawman login API so I can take the format back to the current login page
  • [11:28] Saijanai Kuhn: https://wiki.secondlife.com/wiki/Second_Life_Login_API_Strawman
  • [11:28] Morgaine Dinova: Cool, I'm good at knitting
  • [11:28] Saijanai Kuhn: They did a loto f work but a few bits of punctutation are missing that I'm trying to clear up
  • [11:28] Which Linden: Hey Morgaine
  • [11:29] Morgaine Dinova: waves
  • [11:29] Saijanai Kuhn: BTW, good work on that page. Its so much nicer than what I was doing, I cried quietly.
  • [11:29] Saijanai Kuhn: it being midnight and slightly under the weather didn't help, of course ;-)
  • [11:30] Which Linden: brb, talking irl
  • [11:31] Saijanai Kuhn: OK, I mis-read that. The format looks consistent. Just didn't see the comma before the "--"
  • [11:32] Morgaine Dinova: I missed the start of when things begun to be labelled "strawman" in the project. It's not really the normal use of "straw man". What's the general idea in our context?
  • [11:34] Which Linden: 1 min
  • [11:34] Saijanai Kuhn: well, it started out as a response to Zha's login strawman, I think. And when they first presented it, they were just asking for general feedback
  • [11:35] Saijanai Kuhn: so now it seems to have evolved into the official future login protocol but they haven't changed the name of the page
  • [11:36] Morgaine Dinova: So why was Zha's termed "strawman"? I don't quite relate to the name, since there's no illogical argument involved, afaics.
  • [11:36] CeeLo Linden: declined your inventory offer.
  • [11:37] Saijanai Kuhn: I think she was using it tongue-in-cheek to suggest it was meant to be kicked around because it wasn't really ready to be taken seriously
  • [11:37] Morgaine Dinova: Hmm
  • [11:37] Which Linden: The thing about the strawman is that it describes the current login system,
  • [11:37] Saijanai Kuhn: or something like that
  • [11:39] Saijanai Kuhn: hey Zha
  • [11:39] Saijanai Kuhn: we were just talking about you
  • [11:39] Morgaine Dinova: waves at Zha
  • [11:39] Zha Ewry: blushes
  • [11:40] Saijanai Kuhn: I thought the "strawman" part was a reference to your earlier "login straman" page
  • [11:40] Saijanai Kuhn: https://wiki.secondlife.com/wiki/Second_Life_Login_API_Strawman
  • [11:41] Which Linden: Hey there, back
  • [11:41] Saijanai Kuhn: it wasn't quite ready for beign taken seriously. Now, the new page IS the official page, I think, from what Which says
  • [11:41] Morgaine Dinova: I was just asking why the ame "strawman" is being used. It's not so much a nitpick about the usage of the word being wrong, but more asking what was being implied in our context.
  • [11:41] Which Linden: No, it's not official, per se, I don't think, because, as noted, it's not complete
  • [11:42] Which Linden: Yeah, it seems that "strawman" here is being used like "beta"
  • [11:42] Morgaine Dinova: Ew
  • [11:42] Morgaine Dinova: Well English is an evolving language I guess. :-) Unnecessary overloading though :P
  • [11:42] Saijanai Kuhn: Like I said, I think Zha originally used it as a joke to show that it wasn't meeant to be taken seriously as the end-all of what could be done
  • [11:43] Saijanai Kuhn: at least I THINK it was Zha who set up a login_strawman page at some point
  • [11:43] Zha Ewry: /Newp
  • [11:43] Which Linden: So, is there another strawman page besides the login one?
  • [11:43] Zha Ewry: Oh. Maybe ages ago
  • [11:43] Which Linden: is ignorant
  • [11:43] Saijanai Kuhn: ok, so off in left field or outside the ballpark or some other mixed metaphor
  • [11:43] Zha Ewry: Back before we were re-doing this
  • [11:44] Saijanai Kuhn: I seem to recall Zha doing a brainstorming type page with the "strawman" in the name
  • [11:44] Saijanai Kuhn: so I thought that that was where the name had come from.
  • [11:44] Zha Ewry: sighs
  • [11:44] Zha Ewry: Yes
  • [11:44] Zha Ewry: It did
  • [11:44] Zha Ewry: Ages ago
  • [11:45] Morgaine Dinova: Is the intention to make it document current state of things, and then evolve into something that drives development?
  • [11:45] Saijanai Kuhn: so ancient history aside, I like the format of the page, and the protocols as they are , though more needs to be done, i8'm sure
  • [11:45] Saijanai Kuhn: well, they abreviate my current login page and go from there so...
  • [11:46] Which Linden: Heh, always a good way to go
  • [11:47] Saijanai Kuhn: for my tech-writing demo, I'm going to take that format and revise the current protocol page to fit it. I think the final form will of the new page will need more detail ala what Zha needs to redo things in java
  • [11:47] Which Linden: So, yeah, there is a missing bracket
  • [11:47] Saijanai Kuhn: OK, I wasn't sure. Looking corss-eyed after a while
  • [11:47] Which Linden: Or rather, an extra ]
  • [11:48] Saijanai Kuhn: its pretty much the python form of xml-LLSD, isnt' it?
  • [11:48] Which Linden: Zha, Morgaine, just wanted to trumpet the caps server that Donovan committed last night: [2]
  • [11:49] Which Linden: Yeah, the thing about LLSD is it's not actually a serialization format
  • [11:49] Which Linden: Despite the name
  • [11:49] Which Linden: It's a, uh, data paradigm.
  • [11:49] Saijanai Kuhn: its better than what I'm used to but I've had more COBOL classes than any other language :-(
  • [11:49] Which Linden: So, we use JSON/Python to describe it where that's most appropriate
  • [11:50] Zha Ewry: is mostly AFK, but listenign
  • [11:50] Morgaine Dinova: Sorry, what was that Sai? SL is being rewritten in Cobol?
  • [11:50] Morgaine Dinova: :P
  • [11:50] Zha Ewry: tcl
  • [11:51] Zha Ewry: TCL with extensions
  • [11:51] Morgaine Dinova: I'm tickled
  • [11:51] Saijanai Kuhn: I'm scared (or is it scarred for life?)
  • [11:51] Morgaine Dinova: Hehe
  • [11:51] Which Linden: Hah
  • [11:52] Which Linden: Yeah, so, what else is wrong with the strawman login api>?
  • [11:53] Which Linden: Any troubles in implmentation?
  • [11:53] Saijanai Kuhn: I think its not quite complete enough to implement but I haven't tried yet so...
  • [11:54] Saijanai Kuhn: ah, one thing that is missing, is what format the POST will be in. will it be LLSD-XML?
  • [11:54] Saijanai Kuhn: should have a reference to that and a sample of how it translates, from the protocol format you're using
  • [11:55] Which Linden: I think content negotiation should be used
  • [11:55] Zha Ewry: Pure web approach, when its possible
  • [11:56] Saijanai Kuhn: either way, some discussion of that would be needed before you could implement it from the documentation
  • [11:56] Saijanai Kuhn: wearing the AWG documentor hat here
  • [11:58] Saijanai Kuhn: you've defined the values, but not what format or formats will be acceptable. Python maps, LLSD binary/XML, etc...
  • [11:58] Saijanai Kuhn: or even to determine what formats are acceptable
  • [11:58] Which Linden: Yeah, well we'll certainly support the LLSD serialization formats
  • [11:58] Which Linden: Or at least the XML one
  • [11:59] Which Linden: But, I mean, isn't that what content negotiation means? That you don't have to set in stone what the serialization wil be?
  • [11:59] Saijanai Kuhn: right, so some mention of that should be made plus a simple example and link to the XML-LSLD
  • [11:59] Saijanai Kuhn: sure, but you haven't described how the negoitation takes place, so its still not implementatble as is
  • [12:00] Saijanai Kuhn: this is the kind of thing Zha ad I are supposed to be debating but seldom do ;-)
  • [12:00] Morgaine Dinova: Don't think it would be good to use terms like "Python maps" in a spec that's going to need to be fairly on-the-wire for interop.
  • [12:00] Which Linden: The negotiation should take place via HTTP/1.1 negotiation
  • [12:01] Which Linden: So, uh, we should say that
  • [12:01] Which Linden:  :-)
  • [12:02] Saijanai Kuhn: don't think I'm nitpicking. Zha had to clarify whether we were talking http1.0 or 1.1 to someone based on my original docs
  • [12:03] Which Linden: No, this is great to talk about
  • [12:03] Which Linden: Nitpicking is cleanliness
  • [12:03] Saijanai Kuhn: I didn't even know the difference so I couldn't answer at all
  • [12:03] Saijanai Kuhn: python uses 1.1 by default so it never came up
  • [12:03] Which Linden: Talking to zero now, I think the word is, "using llsd+xml by default, adding more content-types as needed"
  • [12:04] Which Linden: I think that's a good stake in the ground
  • [12:05] Saijanai Kuhn: so thats the default format for both the post and the response...
  • [12:05] Which Linden: Then, in the future, if we decide that we hate llsd+xml, we can migrate to it slowly
  • [12:06] Which Linden: E.g. support both the old and new serializations, then slowly eliminates old-only code
  • [12:06] Saijanai Kuhn: over the next 20 years or so... ;-)
  • [12:06] Morgaine Dinova: We were having a good discussion about content extensibility just after Zero's OH too, in the context of rich IM. The only constant will be change.
  • [12:07] Which Linden: My prediction..... PAIN!
  • [12:07] Morgaine Dinova: lol
  • [12:08] Which Linden: Rome is a big Rocky fan, got me into quoting it at least.
  • [12:08] Saijanai Kuhn: https://wiki.secondlife.com/wiki/Second_Life_Login_API_Strawman#Conventions so should I add something about XML Rpc to that section or to a different section?
  • [12:09] Which Linden: What about xml-rpc? That we currently use it?
  • [12:09] Saijanai Kuhn: xml-LSD sigh
  • [12:09] Saijanai Kuhn: LLSD*
  • [12:09] Saijanai Kuhn: xmlrpc on the brain since I'm editing the current login page
  • [12:10] Saijanai Kuhn: something like "the default format for both POST and response will be in the <link|XML_LLSD> format
  • [12:11] Zha Ewry: The default yes
  • [12:11] Which Linden: OK, well, that might not go on the "login strawman" page, since it's mostly just a description of the current login setup, I think
  • [12:11] Saijanai Kuhn: [[<link>|XML-LLSD format]]
  • [12:11] Which Linden: I don't think there's a document that describes the way login will go in the future, amirite?
  • [12:12] Saijanai Kuhn: well... hmmm. YOu do go into some detail about the new one there though. I think that thats the original purpose of tha tpage
  • [12:12] Which Linden: Bah, it's all confused
  • [12:12] Zha Ewry: Chuckle
  • [12:12] Zha Ewry: Well, that's a good thing to fix
  • [12:13] Zha Ewry: Also.. how publically formal is the XML-LLSD format?
  • [12:13] Which Linden: oic, there's "legacy login" and, uh, not-legacy I guess
  • [12:13] Which Linden: Zha: pretty formal, yes?
  • [12:13] Saijanai Kuhn: This is the page I whipped up for Rob *trying* to clarify things: https://wiki.secondlife.com/wiki/Login_Protocol
  • [12:13] Which Linden: So, yeah, sai, I take it back, the first part of the strawman doc is the right place foor xml-llsd link
  • [12:14] Which Linden: Er, and by "formal" I assumed you meant "calcified and not gonna change"
  • [12:14] Which Linden: No idea how well it's documented, tho
  • [12:14] Zha Ewry: Hah
  • [12:15] Zha Ewry: Well, Suppose i wanted to do a java parse for it?
  • [12:15] Morgaine Dinova: Well if you want formal, let's have a yacc/lex grammar parser for it. Pretty noddy to do.
  • [12:15] Which Linden: The LLSD page seems to conflate all three serialization forms
  • [12:15] Zha Ewry: Not a BNF?
  • [12:15] Morgaine Dinova: Yacc is pretty much BNF
  • [12:15] Saijanai Kuhn: Painini-BNF thanksverymuch
  • [12:16] Which Linden: Hm, something to think about
  • [12:16] Which Linden: The Python LLSD parser is probably the cleanest to read if you want to start with something
  • [12:16] Zha Ewry: Is that out in fre license?
  • [12:16] Saijanai Kuhn: its on the wiki. Think its in the CC license
  • [12:17] Zha Ewry: k
  • [12:17] Which Linden: Hm, the python parser is gpl it seems
  • [12:17] Saijanai Kuhn: ack
  • [12:17] Which Linden: no, wait, it's mit
  • [12:17] Saijanai Kuhn: that really should be changed. i've been referring it to people with the python code I put oup on the wiki which IS in CC license form
  • [12:18] Saijanai Kuhn: np then
  • [12:18] Which Linden: saved!
  • [12:18] Which Linden: [3]
  • [12:18] Which Linden: ok, I gotta go, I'm afraid
  • [12:18] Zha Ewry: excellent
  • [12:18] Saijanai Kuhn: was great chatting which. WIll make those changes on teh strawman page and the current login page
  • [12:18] Zha Ewry: Thanks which
  • [12:19] Which Linden: Thanks for looking the page over, sai, I really appreciate it.
  • [12:19] Which Linden: Peace!