User:Which Linden/Office Hours/2008 Apr 24

[11:03] JayR Cela: yeah I'll stick around , so which can see this
[11:03] Arawn Spitteler: Your browser might be trying to use an Imposter for Baking. Hi, Which
[11:03] Which Linden: Good morning!
[11:03] Saijanai Kuhn: hey all
[11:04] Arawn Spitteler: Has notice been taken of today's upgrade?
[11:04] JayR Cela: hey there Which / long time no see/ and you cant see me / cause my Avatar just went invisible
[11:04] Which Linden: Yeah, that appears to be the case
[11:04] JayR Cela: it happened when I detached a HUD
[11:04] JayR Cela: just poof /
[11:04] Which Linden: You using RC3?
[11:04] JayR Cela: nope
[11:04] Which Linden: :-O
[11:04] Arawn Spitteler: When I tried to log on, at first, I got the dazzle First Look, by mistake, and it tried to upgrade me to a deprecated RC. RC3 isn't working either
[11:05] JayR Cela: oficial 1.19.xxxx with the nicholaz patch
[11:05] Arawn Spitteler: 1.19.0 or 1.19.1?
[11:05] JayR Cela: let me check
[11:05] Arawn Spitteler: Must be hte Windlaugh version, because you're invisible
[11:06] JayR Cela: 1.19.1(4) production release
[11:06] Arawn Spitteler's: dissappointed, that Nicolaz didn't upgrade the 1.19.0
[11:06] Tao Takashi: Hi
[11:06] JayR Cela: and Nicholaz Beresford patch
[11:06] Arawn Spitteler: Hi, Tao, we're admiring JayR's fine form.
[11:06] JayR Cela: and yes Nich did patch this particular Client
[11:06] Tao Takashi: nice nice :)
[11:06] Arawn Spitteler: Hace you tried turning Imposters off?
[11:06] JayR Cela: he says is his last SL client patch though
[11:07] JayR Cela: they are off
[11:07] Tao Takashi: yeah, LL sometimes is good at chasing people away ;-)
[11:07] Arawn Spitteler: Then, my explanations are ended, and the adventures begin
[11:08] Arawn Spitteler: hopes JayR can stay invisible for Torley
[11:08] Saijanai Kuhn: some of us chase after LL but LLL seems uninterested :-/
[11:08] JayR Cela: well I know Nicholaz pretty well / he is very busy in RL / and also he got tired of seeing his fixes be either ignored / or re-broken again & again
[11:08] Which Linden: Yeah, we have to do a better job of talking back
[11:09] Which Linden: My understanding was, many of Nicholaz's patches were too hard to integrate for QA reasons, so we should have told him that, and suggested improvements.
[11:09] Arawn Spitteler: A Nicolaz Liason should be able to explain the patches to LL
[11:09] Tao Takashi: I think it's most of the times communications
[11:09] JayR Cela: so is a combination /of he tired of beating his head against the wall & and he is to busy to be bothered with it / when his work seems to goto naught
[11:09] Tao Takashi: be it this problem or TM or one of the problems before
[11:09] Tao Takashi: anyway...
[11:09] Tao Takashi: Which you should come to EuroPython, time to do a talk proposal! :)
[11:09] Tao Takashi: [1]
[11:10] Which Linden: I'd like to give a talk, *after* CHTTP is in production, but not before.
[11:10] Tao Takashi: then put it in production :)
[11:10] JayR Cela: I gonna re-loog / see if this invisible thingy clears up
[11:10] JayR Cela: brb
[11:10] Which Linden: K, see you in a bit
[11:10] Tao Takashi: some talk about eventlet etc. might be nice
[11:10] Morgaine Dinova: Oooh, people
[11:10] Arawn Spitteler: wants to see you invisible at Torleys
[11:11] Tao Takashi: I will probably do something about AWG
[11:11] Which Linden: I learned at pycon that eventlet isn't polished enough for true evangelism
[11:11] Which Linden: It's good, but it's quite rough and undocuments
[11:11] Tao Takashi: that seemed to have stopped other libs to be ready for prime time ;-)
[11:11] Arawn Spitteler: What's EventLet?
[11:11] Tao Takashi: did you make it an egg in the meanwhile btw?
[11:11] Saijanai Kuhn: contemplates the status of documentation at Linden Labs
[11:11] Which Linden: Arawn: [2]
[11:12] Saijanai Kuhn: Lab*
[11:12] Which Linden: Tao: yeah, exactly
[11:12] Tao Takashi: I now learned to love buildout, very nice to create sandboxes for developing in Python
[11:12] Morgaine Dinova: Reading back to see what I missed
[11:13] Which Linden: We have some refactoring that we need to integrate with the trunk, and then we'll do another beta and maybe an egg
[11:13] Arawn Spitteler: Your invisible neighbor just reappeared
[11:13] Tao Takashi: do an egg now, it's easy ;-)
[11:13] Tao Takashi: or give me access and I will make one ;-)
[11:13] Tao Takashi: actually I did for me python egg tutorial
[11:13] Tao Takashi: me=my
[11:14] Which Linden: We don't want to do an egg now, the API would change.
[11:14] Which Linden: We want to not totally screw over early adopters
[11:14] JayR Cela: yeeaaa / I am back and in solid form again
[11:14] Which Linden: Nice
[11:14] Tao Takashi: increase the version number then :)
[11:15] Arawn Spitteler: doesn't know what a Python is, either: Did you catch, that Dazzle First Look is demanding upgrade to a deprecated RC? Early Adopters volunteered for Screwing over, didn't they?
[11:15] Tao Takashi: because then it's even easier to depend on that egg as you can say "give me version <0.999"
[11:15] JayR Cela: hey is it my imagination / or is there a noticable decrease in UDP protoac traffic recently ?
[11:15] Tao Takashi: needs to write a buildout tutorial
[11:15] Which Linden: Oh, yeah, hm, I'll ping the release team about that arawn
[11:15] JayR Cela: UDP packet
[11:16] Saijanai Kuhn: JayR a lot of the pacets have been converted to tcp
[11:16] Tao Takashi: I guess your av was in such an UDP packet and it got lost ;-)
[11:16] Arawn Spitteler: Accidental Click, on top of their current chaos
[11:16] JayR Cela: about god darn time too
[11:17] Morgaine Dinova: I had an SL packet counter script somewhere, wonder where I put it.
[11:17] Saijanai Kuhn: btw, if anyone could give me feedback on this: https://wiki.secondlife.com/wiki/Current_login_protocols I'm trying to create a "great" example of documentation for Rob L.
[11:17] Tao Takashi: nevertheless helps to see some LL people at EuroPython :)
[11:17] JayR Cela: another question please
[11:17] Tao Takashi: I would like to sprint with them on AWG ;_)
[11:17] JayR Cela: why and the heck arnt all textures compressed to .jpg format
[11:17] Which Linden: Actually, Sai, is there any way to take a look at what percentage of messages have been moved to tcp?
[11:17] JayR Cela: would save a heck of a lot of bandwith
[11:17] Which Linden: JayR: they are, jpg2000
[11:18] JayR Cela: ooops / silly me :_))))
[11:18] Saijanai Kuhn: I wouldn't know how to compared now, with before.
[11:18] JayR Cela: ok then how come thety still take so dammn ling to load then
[11:18] Saijanai Kuhn: I guess, by checking the stuff coming in through event queue get via libsl...
[11:18] Morgaine Dinova: Did lossless jpeg2000 ever get done? I remember there was pressure for it from scultie peeps.
[11:18] Tao Takashi: Sai: You should document the response more
[11:19] Tao Takashi: the non-optional part. This is now an example
[11:19] Saijanai Kuhn: the required part? Heay
[11:19] Saijanai Kuhn: yeah
[11:19] Saijanai Kuhn: wow 4-letter reversal. I'm awake
[11:19] JayR Cela: Morgaine / there is no such thing a lossless in reality with any form of compression / correct me if I am wrong please
[11:19] Morgaine Dinova: You're wqrong :-)
[11:19] Which Linden: I'm pretty sure there is lossless compression
[11:20] Tao Takashi: Sai: with type etc.
[11:20] Saijanai Kuhn: JayR plenty of compression is lossless.
[11:20] Which Linden: gzip for example is lossless
[11:20] Tao Takashi: ZIP is kinda lossless ;-)
[11:20] Morgaine Dinova: Indeed
[11:20] Tao Takashi: I wouldn't like it to lose stuff ;-)
[11:20] Which Linden: also, GIF and PNG are lossless
[11:21] JayR Cela: there has to be some sort of loss / it is impossible for there not to be
[11:21] Morgaine Dinova: And jpeg2000 has its own lossless form too, but it just wasn't being used then since it didn't compress so well.
[11:21] Saijanai Kuhn: like last_name: <string> ?
[11:21] Tao Takashi: JayR: No, it's just encoded differently
[11:21] Morgaine Dinova: Jay: wrong again. There doesn't need to be loss, because most forms of data have a lot of redundancy in them.
[11:21] Tao Takashi: e.g. if you have 10 spaces after another you can say 10 times " "
[11:21] Which Linden: Most compression is about removing redundant information
[11:21] Tao Takashi: this is the most easy way of doing compression
[11:22] Morgaine Dinova: As Which says
[11:22] Saijanai Kuhn: run-array compresion
[11:22] Which Linden: Morgaine: jinx
[11:22] JayR Cela: ok / then how much time and bandwith does it take for the logrithm to reassemble to the origional state
[11:22] Tao Takashi: see also huffman: [3]
[11:22] Morgaine Dinova: Heh. Talking about jinx, is all the land here really grey?
[11:22] JayR Cela: mabey that is why textures take so long to load
[11:22] Arawn Spitteler: Then you have data prioritization, where you decline to download the watermarking of textures
[11:22] Which Linden: Looks normal to me -- perhaps you lost a UDP packet
[11:22] Morgaine Dinova: Hehehe
[11:23] Tao Takashi: it's most likely not the compression which is slow
[11:23] Which Linden: There was a project to do al texture downloads over HTTP, but it got shelved for some reason
[11:23] Tao Takashi: It's LL's new throttle capability ;-)
[11:23] JayR Cela: back to the evil UDP packet protocall / get rid of it / geeze o petes is like kicking a dead horse
[11:23] Morgaine Dinova: I've been sitting here for a couple of hours, so any lost packet should have arrived by now.
[11:23] Tao Takashi: Which: I would like to know the reason :)
[11:23] Which Linden: Ah, but "reliable" packets can fail to be delivered.
[11:24] Tao Takashi: and I also would like to be able to use external textures
[11:24] Which Linden: I don't know, Tao. I'm sure it will rise again.
[11:24] Which Linden: I think "grid is dying, no time for features" is always a reasonable reason.
[11:24] JayR Cela: yes / BUTTTTT !!!!!! UDP has nothing built into the protocal to check for lost packets
[11:24] Tao Takashi: I thought it might stop the dying ;-)
[11:24] Arawn Spitteler: thought the Sim logged out, just an hour ago: You didn't get logged out, Morgaine?
[11:25] Saijanai Kuhn: Which, when did that hapen? Zero was teling us that textures were 2x faster during testing.
[11:25] Tao Takashi: I never log out, that feels like being dead
[11:25] Which Linden: Tao: I think if you downloaded textures more quickly, it might place more load on the asset server.
[11:25] JayR Cela: mabey with this new sever code they rolling out 2day / perhaps
[11:25] Which Linden: Sai: Zero would know more, but it was almost a year ago.
[11:25] Tao Takashi: Which: Another reason to make them load from external.. then they still load slow but it's not LL's fault anymore ;-)
[11:25] Which Linden: Ha ha, yeah
[11:26] Tao Takashi: and it will encourage people to build better asset servers ;-)
[11:26] JayR Cela: all in all I do see improvements
[11:26] Morgaine Dinova: Well, one can always answer the feature advocates "Yes, we did implement the feature you wanted, but it isn't showing because of the bugs which we couldn't fix because we were working on your feature."
[11:26] Tao Takashi: ok, so as long as there's nothing wrong with http textures in general I am fine
[11:27] JayR Cela: i think SideWinder / Pastrami/ prospero/ joshua / are excellant examples of how things shoud be dealt with
[11:27] Which Linden: Yeah, those guys are awesome
[11:27] Tao Takashi: I guess I always miss their OHs ;-)
[11:27] Tao Takashi: and I guess this google calendar full with invalid data is one of the reasons I do ;-)
[11:27] JayR Cela: you aint kidding / 2 and a half years I been in-world / never recall such a level of professionalism
[11:28] Saijanai Kuhn: google seems to have problems with daylight savings, at least for me
[11:28] Which Linden: Don't you know it's a beta calendar? >:-)
[11:28] JayR Cela: Google is like SL / perpetual Beta
[11:29] JayR Cela: lol :_)))
[11:29] Tao Takashi: it's not a beta calendar but just outdated ;-)
[11:29] JayR Cela: anyways aside for a few pesky annoinces / things seem to be working fairly well for me
[11:30] Which Linden: Yay good to hear
[11:30] pegaze Barbasz: hello
[11:30] Tao Takashi: Hm, the OpenSocial people are now talking about a PRESENCE proposal...
[11:30] Arawn Spitteler: thought Invisibility looked fun
[11:30] Tao Takashi: I guess they have different requirements than we have..
[11:30] Which Linden: Yeah, their presence is probably pretty static
[11:30] JayR Cela: I just downloaded the latest Nvidia 174.93 driver and got the tweaked version from ??????
[11:30] Tao Takashi: not necessarily, they also want to know who of your friends is online
[11:30] JayR Cela: and it runs great
[11:31] Tao Takashi: so an opensocial gadget can show that
[11:31] Which Linden: Oh, hah, their presence is just a flag.
[11:31] pegaze Barbasz: stop
[11:31] Which Linden: hammer time
[11:31] Tao Takashi: so what is it in SL?
[11:31] Tao Takashi: to me it looks like either on or not ;-)
[11:31] Arawn Spitteler: Is the latest NVidea Driver supported, now?
[11:31] Which Linden: It's an xml document describing your location.
[11:32] Which Linden: And some other data like state.
[11:32] JayR Cela: this last one is not for SLI rigs
[11:32] Which Linden: (so, a flag, plus a url, basically)
[11:32] Tao Takashi: ok, but when I know somebody is on can't I do then a separate query for getting the location etc.?
[11:32] Tao Takashi: well, that might be some URL
[11:32] Which Linden: Yeah, I guess.
[11:32] JayR Cela: hey I give you the best place for ATI & Nvidia drivers
[11:32] JayR Cela: [4]
[11:32] Which Linden: But, you know, might as well just do a third query to see if they're online or not
[11:32] Tao Takashi: if we think about presence in the DataPortability Group it might also be something with location
[11:33] Tao Takashi: so on which social networks you are online that is
[11:33] Tao Takashi: or on which
[11:33] JayR Cela: they allways 2 steps ahead of the stock drive releases
[11:33] Tao Takashi: so we will see if some standard arises from that. If it does then it might be nice to support it with SL, too
[11:33] Tao Takashi: most likely in an AWG based one
[11:34] Morgaine Dinova: Tao: I just posted to a friends' list your line "I never log out, that feels like being dead" --- they liked it ;-)
[11:34] Tao Takashi: heh :)
[11:34] Arawn Spitteler: Death is just nature's way, of logging us off.
[11:34] Which Linden: Verily
[11:34] Tao Takashi: it's just a flag ;-)
[11:34] Morgaine Dinova: Death is just an engineering problem.
[11:35] Tao Takashi: might be a bug
[11:35] Which Linden: So, does anyone want to continue the conversation from last time about HTTP errors?
[11:35] Arawn Spitteler: Don't blame me, for your limited life span, unless you want to agree with Genesis.
[11:35] JayR Cela: looks as if I am about to crash
[11:35] Tao Takashi: I guess I missed last time
[11:35] JayR Cela: brb
[11:35] Morgaine Dinova: Yep, sounds good.
[11:36] Arawn Spitteler: hopes to see her invisible again
[11:36] Morgaine Dinova: We need a recap though first :-)
[11:36] Which Linden: Yeah, OK, and since I failed to put up the transcripts
[11:36] Morgaine Dinova: Hehe
[11:36] Morgaine Dinova: Tut tut, no waterinmg the bamboo for that
[11:36] Which Linden: It's such a pain to do so. Cash money to whoever comes up with a web service that's just cut-n-paste
[11:37] Morgaine Dinova: Yeah. It should be so simple to keep a logger object on the parcel, with simple record/stop/publish buttons
[11:38] Arawn Spitteler: It should be doable, with a llistener, using llHTTPRequest() shouldn't it?
[11:38] Which Linden: Should be
[11:38] Tao Takashi: I might do a Plone based logger this weekend ;-)
[11:38] Which Linden: So, from last time, "So, the example we had, the web service that transfers L$ from one agent to another returns with a 4xx if you leave out a parameter, but a 200 if the source doesn't have enough of a balance for the transaction to succeed. "
[11:38] Morgaine Dinova: k
[11:39] Arawn Spitteler: Should it start recording, for an hour, if llDetectedKey() is llSameGroup()?
[11:39] Which Linden: Since that time, we actually changed it so it doesn't even return a 4xx when the parameters are wrong
[11:39] Which Linden: It now only returns 4xx if the url is wrong
[11:40] Morgaine Dinova: So in other words, you're decoupling service error semantics from HTML error semantis.
[11:40] Which Linden: *HTTP error semantics
[11:40] Morgaine Dinova: Good correction, yep
[11:40] Which Linden: Yeah, so basically once it reaches our application code it doesn't return anything but 200
[11:41] Arawn Spitteler: It doesn't return an "Excellent Command or File Name!?"
[11:41] Morgaine Dinova: OK. Which of course begs the question of how service-level error conditions are sensed.
[11:42] Which Linden: What do you mean, service-level?
[11:42] Which Linden: Like if the web server is down?
[11:42] Which Linden: Or if the database backing it is down?
[11:42] Morgaine Dinova: No, like when parameters are wrong.
[11:42] Morgaine Dinova: Or even acocunt empty
[11:42] Morgaine Dinova: account*
[11:43] Which Linden: Those are currently 200s
[11:43] Which Linden: We have a standardized formate for returning errors
[11:43] Morgaine Dinova: Oh, each one has a unique 2xx ?
[11:44] Which Linden: It's an llsd struture like {'error':{'message':'I broke', 'identifier': 'broken'}}
[11:44] Which Linden: Only 200, the body differs tho
[11:44] Morgaine Dinova: Ah OK. That's fine
[11:44] Which Linden: The identifier can be used to determine the nature of the error, the message is human-readable
[11:45] Which Linden: Which is really good because we went through the code, and the error messages I wrote used "can't", and the ones jwolk wrote used "cannot"
[11:45] Morgaine Dinova: So the semantic of the HTTP transport is "If there is a valid LLSD returned, then the transport operation has succeeded" --- including when the LLSD is an error structure.
[11:45] Which Linden: Yes, I think that is the idea
[11:45] Arawn Spitteler: welcomes Lucifer's Loss: Hi, April
[11:46] Aprill Hellman: Hi... how are you?
[11:46] Morgaine Dinova: That sounds clean enough.
[11:47] Which Linden: Downside is, you can't look at the logs and see how many requests are truly succeeding
[11:47] Tao Takashi: you might need different logs then
[11:47] Morgaine Dinova: But that's good, you should be seeing zero failures in the web logs, otherwise there's a client programming problem.
[11:48] Arawn Spitteler: doesn't think there could be any client programming problems: You have Client Programming Adventures, and Server Programming Problems.
[11:49] Which Linden: Yeah, you could even set up triggers that would page you when the number of non-2xx codes crossed a threshold
[11:49] Morgaine Dinova: Whereas not having enough L$ for a transaction should not be something that a web-level admin should worry about. It's an error condition at the transaction level, not that the web infrastructure level.
[11:49] Which Linden: Yup
[11:50] Which Linden: I'm worried though that we'll end up making everything have 2xx error codes
[11:50] Morgaine Dinova: Sounds clean then
[11:50] Which Linden: I'm trying to come up with examples of where you'd want it to give a 4xx or 500
[11:50] Which Linden: Obviously, a 500 is given for every uncaught exception
[11:50] Arawn Spitteler: Something's occured to me, of permissible bank errors. Banks have to carry insurance, and still lose money, in an average transaction, both of which have to be covered by Profit Margin, to remain viable. Perfection would sink the Profit Margin.
[11:51] Which Linden: A 403, I guess, if you don't have permission to perform an operation.
[11:51] Which Linden: Arawn: really? banks just carry a load of money to cover for mistakes?
[11:52] Arawn Spitteler: Banks have to earn a load of money, to cover those mistakes. It's amatter of how many mistakes optimize viability
[11:52] Arawn Spitteler: Small mistakes can be covered out of a teller's income.
[11:53] Which Linden: I guess that makes sense. However, the world of a bank is much more wooly than the world of Second Life.
[11:53] Arawn Spitteler: Not if Linden Research Labs is the Bank of Second Life.
[11:53] Which Linden: (despite the fact that SL seems to be the wooliest place in the universe sometimes)
[11:53] Morgaine Dinova: A VW transaction system will have its "banking errors" too, failures of infrastructure or programming bugs, it's inevitable. I wonder how that is going to be handled.
[11:53] Arawn Spitteler: SL is an immersive wooliverse.
[11:53] Which Linden: LOL
[11:54] Which Linden: Yeah, that's true, Morgaine. I don't know what we do for those right now
[11:54] Tao Takashi: I gotta go, RL calls (quite distracted already anyway).. cya later!
[11:54] Which Linden: Presumably we will simply have to have humans investigate all errors, and just try to keep them at an acceptable level, as banks fo
[11:54] Which Linden: Thanks for stopping by Tao
[11:54] Morgaine Dinova: I doubt that LL wants to be the bank of the SL wooliverse. The feds are enough trouble to Philip already.
[11:54] Tao Takashi: thanks for the meeting! :)
[11:55] Arawn Spitteler: has been trying to re-market The Holy Grail, but the bleeding edge nature of LRL suggests the Pillum of Longenous might be more appropriate.
[11:55] Morgaine Dinova: Cya Tao
[11:56] Which Linden: Yeah, I don't know about "bank", but for the forseeable future we'll certainly be a transport mechanism
[11:56] Arawn Spitteler: Acting as acountant of L$ is going to make for some degree of Banking.
[11:57] Ellla McMahon: thank you ... see everyone soon :)
[11:57] Which Linden: Seeya ella
[11:57] Morgaine Dinova: Which: with a "perfect" transport and transaction mechanism, you don't really need to be a "bank", in Arawn's sense --- no insurance needed if nothing fails. Unfortunately, reality's not like that. There will be trouble ahead.
[11:58] Which Linden: Yeah
[11:58] Which Linden: Even if the technology works perfectly
[11:58] Arawn Spitteler: "Learning Curve" Will cost more than Time.
[11:58] Which Linden: There will be human failures, as in "L$ didn't arrive in time so my contract was broken, etc etc"
[11:59] Arawn Spitteler: Lindens will be dealling with Problems, as will other users.
[11:59] Morgaine Dinova: Which: yeah. You'll need some realtime programming to ensure transaction occur within realtime set limits or not at all.
[12:00] Arawn Spitteler: Helping Bankers share in the getting of the wisdom of their ministry, is what I'd suggest. It might also help, to find a past life heir of Dagda, and a past life Templar
[12:00] Arawn Spitteler: Compare Credit Cards to Overdraft Protection, and what's a Letter of Credit?
[12:01] Morgaine Dinova: I guess there will have to be an SLA from LL for financial transaction at some point. In litigious USA, I can see it becoming necessary./
[12:01] Which Linden: I'm afraid I don't understand what you're saying, Arawn.
[12:01] Which Linden: What's an SLA?
[12:01] Morgaine Dinova: Service Level Agreement
[12:02] JayR Cela: wow that was odd
[12:02] Arawn Spitteler: Something better than, "We're right, and the Final Arbiters of Justice."
[12:02] JayR Cela: could not logg-in for a while
[12:03] JayR Cela: so what did I miss
[12:03] Which Linden: We talked about HTTP error codes, proper use of.
[12:03] Arawn Spitteler: Welcome back, JayR. I'm talking about making my fortune, resurrecting Magic, as the foundations of Guilds. Dagda's Cauldron and The Holy Grail have reall places in Economic History
[12:03] JayR Cela: or lack there-of :_)))))
[12:04] Which Linden: Morgaine: we have a lot of reliability work to do before we can support an SAL
[12:04] Which Linden: *SLA
[12:04] Morgaine Dinova: An SLA for financial transactions would probably cover availability levels and transaction times at the very least, plus the remedies available when not met.
[12:04] Morgaine Dinova: Yes indeed
[12:04] Which Linden: Ah, it's after noon. I should mosey on in to the office.
[12:05] JayR Cela: i have one last comment
[12:05] JayR Cela: please
[12:05] Which Linden: Whuzzat
[12:05] Arawn Spitteler: Aye, M and Phillip and Zero, have got to get togeather with some Bankers, in the study of History.
[12:05] JayR Cela: someone should explain to whom ever what exactly a release candidate actually is
[12:05] Arawn Spitteler: Royal Clown Viewer?
[12:06] JayR Cela: these last half dozen RC's are not release candidates they are late Alpha early beta at best
[12:06] Morgaine Dinova: A release candidate is code that escaped before someone could close the door. ;-)
[12:06] Which Linden: Ha ha, yeah. Some could say that the Release client is late beta itself.
[12:06] Arawn Spitteler: 1.20.3 is going to be availble, before 1.20.4, or not.
[12:06] JayR Cela: they are not ready to be put out to the average Avatar
[12:07] Which Linden: No, and the average avatar shouldn't be downloading RCs
[12:07] JayR Cela: they are garbage
[12:07] Morgaine Dinova: So why use them Jay?
[12:07] Which Linden: They are for early adopters.
[12:07] JayR Cela: well your correct Which / but the fact.....
[12:07] Which Linden: And those with a taste for pain.
[12:07] Arawn Spitteler: The Average Client should be using RCs, but for unimportant functions.
[12:07] JayR Cela: that LL makes them available to the general Avatar is a BIGGGGGG mISTAKE
[12:07] Morgaine Dinova: I have no taste for pain, and so have never tried an RC. Simple ;-)
[12:07] Mikela Low: hi girls
[12:08] Which Linden: Dunno, the RCs are kinda hidden on our download site, and I think the blog readers are hardly "average"
[12:08] JayR Cela: i think / the RC's should be moved back to the Beta Grid
[12:08] Morgaine Dinova: When I want pain, I'll roll my own clients. Then it can be guaranteed pain ;-)
[12:08] Arawn Spitteler: smiles at Mikela, and takes a moment to look above her legs: Hi, Mikela. Some of us were just heading late to Torley's
[12:08] JayR Cela: and not available to the general population
[12:09] JayR Cela: and then add a perk of somesort to the folks willing to actually help with beta testing
[12:09] JayR Cela: currently there is no real reason for folks to want to help out
[12:09] Arawn Spitteler: Beta Test Montors should have special access to free tech support
[12:10] Morgaine Dinova: They already get the main perk, ie. pain. RCs are for ackowledged masochists ;-)
[12:10] JayR Cela: and then these crappy RC's are let loose into the void
[12:10] Which Linden: Well, that's an alternate poiint of view, but the problem with the beta grid was that it wasn't real enough, people weren't doing their normal activities and thus not finding bugs that existed.
[12:10] Arawn Spitteler: Worse, the RCs become the Main Viewer, without being tested.
[12:10] Which Linden: Sure they're tested.
[12:10] JayR Cela: well , Which at this point the RC's are seen by anyone that has been here a while as a complete joke
[12:11] Which Linden: If the RC is out in the wild for a few days and no new bugs are found, it's blessed as release
[12:11] JayR Cela: problem is the Joe Blow average person does not know any better
[12:11] Morgaine Dinova: I guess you just need more human bodies than now in QA, to weed out the worst.
[12:11] Which Linden: Yeah, we're hiring.
[12:11] Arawn Spitteler: wishes he had a resume.
[12:12] Arawn Spitteler: How much do they pay, in QC?
[12:12] JayR Cela: either that or put a stearner warning label on the RC's before download access is granted
[12:12] Which Linden: It is a bummer that old bugs aren't really considered to be blockers in the release process, but we'd never fix anything at all if that were the case.
[12:12] Which Linden: Arawn: no idea
[12:12] Which Linden: JayR: the splash page for the RC is pretty loud about the quality
[12:12] Morgaine Dinova: Which: mind you, it's a problem of your own making. You "opened the client" allegedly, yet disallow commnity SVN access, and you don't build and release a parallel "community client". So you have problems.
[12:13] JayR Cela: not loud enough Which
[12:13] Which Linden: Yeah, we should be doing that
[12:13] Which Linden: Completely agreed.
[12:13] Arawn Spitteler: SVN?
[12:13] Which Linden: We are waiting for some projects to come through before we can do that though.
[12:13] Which Linden: IT would take too much time on our current platform.
[12:13] JayR Cela: needs to be / like hit you in the head with a brick / if you dont know WTF you are doing or wish to help contribute to futher progress / DONT DOWN LOAD this
[12:13] Morgaine Dinova: You'e not harnessing the community. And as some have pointed out, you're even pushing some devs away, like Nicholaz.
[12:14] Morgaine Dinova: Aye
[12:14] Which Linden: I wish we could figure out a way to harness the community in some way that didn't involve patches.
[12:14] Which Linden: We kind of are in the form of RC and PJIRA
[12:14] Which Linden: Kind of extended QA
[12:15] Which Linden: That's an important factor that most people overlook
[12:15] Which Linden: Building a product involves much more than simply code.
[12:15] JayR Cela: yeah and no-one actually gets paid for their time / thats why no-one gives a crap / all they do is bitch
[12:15] Which Linden: Ha ha, well, there is a certain amount of noise in any system.
[12:15] JayR Cela: pay em 500 lindens to test
[12:15] Arawn Spitteler: Forums could stand to be opened. Nicolaz is an example. LRL should develop an algorythm where investors feel satisfied, and not all investors have dollar value expectations
[12:16] JayR Cela: just my opinion...........
[12:16] Morgaine Dinova: Which: Parallel streams, any number of them. Continue doing your own "official" client, but provide SVNs for several other streams (you can contribute to them too, but don't *regulate* them). Then at release time, you build the entire set of clients, and publish them with appropriate disclaimers. If some don't build, that's not your problem.
[12:17] Arawn Spitteler: Andrew once offered a chunk of change, for the first ten people who could crash Crash Me, and Viewer Crash Prizes might be a good bounty
[12:17] Which Linden: Well, others can set up their own svn servers, Morgaine, I don't see how it's any better if we host the svn server.
[12:18] Morgaine Dinova: Which: the important thing is, YOU BUILD and PUBLISH them. Currently, you're ignoring community devs --- that's the main criticism, and problem.
[12:18] Which Linden: In fact, it's bad for us to host the svn server because no matter how many disclaimers we put on it, we will still be blamed.
[12:18] Which Linden: We're not ignoring them completely. We just don't have 100% attention
[12:18] Which Linden: Not everyone can be heard, we have to pick and choose.
[12:18] JayR Cela: well their are a number of reasons for that problem........
[12:18] Which Linden: But I personally read every email sent to sldev, all the way through.
[12:19] Morgaine Dinova: Which: well, nincompoops who choose not to read what they're downloading will complain about LL, sure. But they're nincompoops. Ignore them ;-)
[12:19] JayR Cela: one of them being most people do not bother to read disclaimers / or instruction manuals for that matter either
[12:19] Morgaine Dinova: Jay: that's their problem.
[12:19] Which Linden: Yeah, well, um, then there's more noise and we'll overlook real developers because they got stuck in with the nincompoops
[12:19] Arawn Spitteler's: brother had a story, about the Microsloth Readme File
[12:20] JayR Cela: ok here is an idea
[12:20] JayR Cela: make it so there is a verifacation that the disclaimer and release Notes have been read before proceading
[12:21] JayR Cela: and I aint just talking about scrolling down to the bottom of the page
[12:21] Which Linden: Have a little quiz?
[12:21] Morgaine Dinova: Which: the FOSS community is well versed in keeping nincompoops from mucking up the code. Not everybody gets write access to a given SVN repository -- its leader decides. Of course a disgruntled nincompoop can create his own (and you would still build it for fairness), but everyone else will ignore his stream.
[12:21] JayR Cela: some thing like that / yes
[12:21] Arawn Spitteler: A check mark for each possible disaster could sort out which downloads you're allowed?
[12:22] JayR Cela: there ya go
[12:22] Which Linden: Morgaine: yes, I think we need to follow the Linux model in that regard. However, like I said, we'd like to have some tecmology in place before we can arrange that.
[12:22] Morgaine Dinova: Time is passing ... :-(
[12:22] Which Linden: Yah
[12:22] Which Linden: I'll suggest something
[12:22] JayR Cela: that alone should discourage all but the most devoted folks that actually want to help out from the people that continually piss & moan
[12:23] Kristaki Hudson: hello
[12:23] Which Linden: Ah, but, as noted, JayR, there will always be pissing and moaning, nothing you can do about it
[12:23] Which Linden: Hi Kristaki
[12:23] Morgaine Dinova: Yeah, pissing and moaning is not going to stop, lol.
[12:23] Arawn Spitteler: First Look could be renamed "Piss and Moan Candidate."
[12:23] JayR Cela: yeah / but it may get rid of quite a few of the / jack-ass;s
[12:23] Which Linden: Ha ha ha
[12:23] Morgaine Dinova: Hehe
[12:24] Kristaki Hudson: sorry i think u are in conference i go further have a nice evening
[12:24] Arawn Spitteler: Brain Storming, perhaps you'd like to share the suffereing?
[12:24] Morgaine Dinova: Cya Kristaki. Yep, it's a technical meeting, sorry
[12:24] Which Linden: Hey, this has been a gas, but I do have to go.
[12:25] Morgaine Dinova: KK Which, have fun in the office ;-)
[12:25] Which Linden: JayR: I sugegst you go to Bridie's office hours with your RC concerns
[12:25] Arawn Spitteler: If the Gas is that bad, Going is a good idea
[12:25] JayR Cela: ok / which / later / thanks for having us over
[12:25] Which Linden: She is the PM for that thing
[12:25] Which Linden: Laters!

User:Which Linden/Office Hours/2008 Apr 24

Navigation menu

Search