Web   ·   Wiki   ·   Activities   ·   Blog   ·   Lists   ·   Chat   ·   Meeting   ·   Bugs   ·   Git   ·   Translate   ·   Archive   ·   People   ·   Donate

#olpc-meeting, 2009-01-15

Index | Today     Channels | Search | Join

All times shown according to UTC.

Time Nick Message
18:01 mchua_ Ha, there we go.
18:02 I'm not sure if anyone else is around for- oh!
18:02 hi, garycmartin and marcopg!
18:02 marcopg hello mchua_ garycmartin!
18:02 garycmartin mchua_: evening mel!
18:03 mchua_ realizes her client may have not reconnected when she got to 1cc. that would explain a lot.
18:03 garycmartin Hi Marco!
18:03 marcopg will only follow tonight, his head is hurting too much to be useful!
18:03 mchua_ There's not much to go through tonight. http://wiki.laptop.org/go/Comm[…]etings/2009-01-15
18:03 garycmartin, haven't seen you for a while. welcome back. :)
18:03 any additions? else we can make this superfast and be out of here in 10min or so. ;)
18:03 and I'll get back to 8.2.1 testin'.
18:04 # 2 Testing requests
18:04 # 3 Upstream testing
18:04 # 4 8.2.1 test results
18:04 holler with additions and we'll queue them up at the end. :)
18:04 in the meantime, launching into testing requests...
18:05 we've got some activity testing requests for socialcalc, foodforce, and gnubook.
18:05 should we be pushing these upstream to the (brand-spankin'-new) SL Activity group?
18:06 garycmartin * holds his SL Activity group hand up :-)
18:06 mchua_ votes yes, looks around
18:06 marcopg mchua_: is activity team going to do testing?
18:06 mchua_ worships garycmartin
18:06 marcopg: Well, if they're taking responsibility for Activities, I hope so. :)
18:06 garycmartin :-)
18:07 marcopg mchua_: yeah was thinking of a development/testing separation, but perhaps for activities they can be coped
18:07 garycmartin I need to make a nice new icon for socialcalc.
18:07 mchua_ so we'll take that as a yes, I suppose, and - garycmartin, I don't know a whole lot about what the Activity team is up to, but would you mind making sure the migration of Activity testing to the SLAT (...huh, acronyms!) goes over all right?
18:08 mm, icons.
18:08 garycmartin Activity developers should be activity testers... In my world at least, even if it's only oassionally.
18:08 mchua_ nods
18:08 marcopg, I guess the thing we'd have to figure out is what the infrastructure for Activities will be at SL, so we can smoosh all the current docs and test results/cases that-a-way.
18:09 marcopg mchua_: can we just start from a plain dlo -> SL move?
18:09 wlo
18:09 mchua_ marcopg: yup.
18:09 marcopg looks at wikiteam lead david :)
18:10 mchua_ More complicated things are a bernie question, I think, and if garycmartin can make sure the QA angle gets brought up when the Activity team starts talking, then I'm happy and we can transfer the data over when we have a place to dump it.
18:10 marcopg he has bots that can do the move
18:10 we will have to make sure we have all the trac plugins we need I guess
18:10 mchua_ anything else? we can start telling activity authors to look at the Activities list...
18:11 garycmartin Activities list?
18:11 marcopg sounds good to me
18:11 mchua_ er, Activities team.
18:12 garycmartin I think the team is going to start go foraging for devs... best way really.
18:12 mchua_ yay!
18:12 marcopg :)
18:12 garycmartin: (off topic, but you are doing awesome things there!)
18:12 mchua_ anything else on activity testing? I'll email the 3 testing requestors individually and let them know what's going on (I want to actually get each QA group to focus on testing what they should be testing, too.)
18:12 marcopg +1; garycmartin is awesome.
18:13 hullo tabitha!
18:13 marcopg hi tabitha
18:13 tabitha Hello
18:13 sorry I am late
18:13 mchua_ anyhoo, moving on... anything else on that topic?
18:14 no worries, tabitha. :) glad you're here.
18:14 marcopg that's all prolly, but I can't quite think right now :)
18:15 mchua_ tabitha, we just agreed to move activity testing to sugarlabs now that they have an Activity team - but that doesn't change the stuff you and the welly testers are doing at the moment (when the infrastructure migrates, there'll be big announcements to the list.)
18:15 tabitha cool
18:15 mchua_ tabitha: though if you're looking for things to test tomorrow, these three projects requested a look at their Activities. http://wiki.laptop.org/go/Comm[…]#Testing_requests
18:15 moving on!
18:15 upstream testing. http://wiki.laptop.org/go/Comm[…]#Upstream_testing
18:15 short version: we now have it.
18:15 tabitha great - love being given direction on what to test
18:16 mchua_ does anyone here have experience with Fedora QA?
18:16 garycmartin mchua_: sorry, no, XO distro only here.
18:16 mchua_ so basically, since OLPC's software dev efforts just became much smaller, OLPC's QA group's job has become much easier
18:17 ok, I guess that's something that I can rope Jeff, Colin, and Sandra into checking out then. :)
18:17 marcopg mchua_: no exp, other than fixing bugs in their tracker
18:17 mchua_ they all expressed interest in learning about testing, and the Fedora QA group has more scaffolding to guide them, I think.
18:17 marcopg weee less work to do ;)
18:18 mchua_ So I'll rope those folks into finding out what's going on there and if there's anything we need to do downstream...
18:18 our other big upstream is Sugar Labs, and we already moved Activity testing up there...
18:18 and... can finish worrying about testing Sugar itself after the feature freeze when we know what we'll be testing, so that's next week...
18:18 marcopg we will have start tracking Sugar on Fedora bugs at Fedora I guess
18:19 mchua_ nods
18:19 marcopg not sure how to communicate about that, people will keep using dev.laptop.org for a while I guess
18:19 mchua_ so there are 3 big OLPC QA things to keep in mind
18:19 marcopg maybe some announcement + answering d.l.o
18:19 listens
18:19 mchua_ 1) is communication with upstream, which is pretty minimal, and what we're just doing now, and once that's figured out as long as there's a line of communication we can use then everything should be fine - once things have settled so that we know who is doing what.
18:20 2) is tests that require large-scale testbeds of many XOs, an XS, etc.
18:20 that's something I have to keep talking with ed about to see if OLPC can give equipment to test groups to get 50 XO testbeds in, for instance, an engineering college near large deployments
18:20 (engineering colleges having things like spectrum analyzers hanging around)
18:21 garycmartin mchua_: is Joe still at OLPC?
18:21 mchua_ and 3) which I don't know a lot about yet is moving testing towards being done by deployments themselves, particularly for acceptance testing.
18:21 garycmartin: no, we have exactly 0 QA staff now
18:21 garycmartin: (I'm also no longer at OLPC, technically.)
18:21 marcopg 3 sounds pretty cool when we can get that moving
18:22 might be bit complicated to setup perhaps
18:22 mchua_ so 3) is something I'd really like to talk with folks about at LCA, since tabitha and others have such a strong QA presence there *and* there's this awesome deployment stuff going on.
18:22 garycmartin mchua_: Yea I knew about you (blogs et al) had been working with Joe on QC bulk XO testing scripts.
18:22 marcopg sounds good
18:22 mchua_ so I think you guys are probably our best bets to set up a model deployment in terms of "we do QA in a totally awesome local manner!" if that's the direction that we want to go in.
18:22 marcopg is 1 something you are interested to go into detail here?
18:23 doesn't know anything about 2 :)
18:23 mchua_ marcopg: we covered that briefly, a bit, with "Fedora is going to be taken care of by these people who'll check into things, and SL will wait a week" - did you have any particular questions?
18:24 garycmartin: ah, cool - I didn't realize you were still working on that, is there anything we should be picking up on / playing with? ;)
18:24 marcopg mchua_: missed the week bit, make sense
18:25 garycmartin mchua_: well, unless you have 50+ XOs you need to remotly manage and log data from, no...
18:25 mchua_ garycmartin: Not yet, but I would love to have that problem.
18:25 garycmartin mchua_: :-)
18:25 marcopg hehe
18:25 mchua_ It's called the "we have testbeds outside 1cc that are community-accessible!" problem.
18:26 garycmartin: good to know that's waiting in the wings for when we do have that problem, though. :)
18:26 garycmartin mchua_: The 3 I have here is not really enough for bulk stuff ;-)
18:26 mchua_ that's a bit of vague and shaky overview of The New Direction Of OLPC QA (much smaller than it used to be) - things should get clearer in the weeks to come as we continue testing
18:27 any questions on what's going on? I think it'll be much more sorted out once we boot 8.2.1 out the door, honestly
18:27 since that represents the last anomaly of "the old structure" that we have to clear out
18:27 garycmartin mchua_: I wonder if we can run 50+ virtual Sugar machines as a SW test bed.... Hmmm...
18:27 mchua_ and then we can really turn our attention to getting local QA groups at deployments going
18:28 garycmartin: ooo, that would be sweet. I think the tricky thing is that a large part of the big testbeds will be networking.
18:28 marcopg garycmartin: people doing thin client stuff thought that was pretty easy/possible
18:28 but yeah doesn't really cover networking
18:28 mchua_ mesh in different environments, various configurations of APs and XSs, etc - *however* one thing that I haven't seen done yet is making simple networking simulations for that
18:29 so that we can have predictions/suggestions of what things will look like without whipping out 50 XOs.
18:29 garycmartin mchua_: well bandwidth could be throttled to simulate, but yea it's only one step on the road (but a good one if possible).
18:29 mchua_ a great one, if possible.
18:29 marcopg agreed that would be useful
18:29 mchua_ is that something you're interested in looking into, garycmartin? I know you've got a lot on your plate right now.
18:30 I wonder if we can find some engineering undergrad or masters students taking a communications class that could do this as an assignment.
18:30 but I'll leave that on the table in case someone wants to pick it up.
18:31 garycmartin mchua_: I can't emulate 1 Sugar OS here just now, so 50+ is a little hopefull!!! :-)
18:31 marcopg hehe
18:31 mchua_ the big thing on my plate right now is 8.2.1 testing, which I'm getting through tonight (now that I have 2 XOs of my own at home - minitestbed!)
18:31 marcopg haven't seen brendan around lately, he was looking into it
18:31 will ask him about it, when he is around
18:32 mchua_ marcopg: sugar emulation?
18:32 marcopg mchua_: yeah
18:32 brendan was at sugarcamp and we talked briefly about it
18:32 mchua_ Cool.
18:33 garycmartin mchua_: regarding 8.2.1, the change list was hard to test, it's pretty much all deployment level stuff. Thats why I just did a full G1G1 Activity test run though.
18:33 marcopg is there anything I can do to help with 8.2.1?
18:33 mchua_ garycmartin: yeah, that's why I started realizing I might need to do most of the 8.2.1 testing, since we have the testbeds here and all.
18:33 thanks for doing Activity testing, though!
18:33 marcopg would reeeeally like to see it out of the door to get new sugar in the hands of the kids
18:34 mchua_ garycmartin: marcopg: what would help a lot, since you folks are both developers, is to help take a look at the code changes that were made and come up with test cases i should run in areas of possible regression
18:34 tabitha_ do you want us updating all XOs to 8.2.1 or should we keep any on 8.2-767 for continuing any of the g1g1 tests?
18:35 marcopg mchua_: ok, cool
18:35 mchua_ tabitha_: Hm - I don't think it should make a difference to Activity testing, so it's your call. I'd probably save the time and just stay on 8.2-767, unless someone in the group really wants to test NANDblaster or something. :)
18:35 marcopg (will try to look into the sugar related ones)
18:36 garycmartin mchua_: NANDblasting was fun. Worked well, but need to be aware of issues when making clones of an XO.
18:37 mchua_ marcopg: thanks! I think that's just #9044 and #8745 so far (and they're both translations); the rest are lower-level.
18:37 garycmartin: the issues you wrote up in your wonderful email to the testing list?
18:37 marcopg aaah thanks for the numbers that's very useful ;)
18:37 mchua_ I'm particularly in need of folks who can help spot potential areas of networking regression, so I can run them with the Big Testbeds here
18:37 marcopg oh both translations!
18:37 so I can't be very useful, gah ;)
18:38 garycmartin mchua_: Kill list of files that should be removed after a clone. I didn't specifically email about this (other than posting some nandblaster stats to the list from my net environ)
18:38 mchua_ marcopg: that's ok, you have plenty to do on sugar. :) get the activity folks up and running.
18:38 marcopg mchua_: :)
18:39 mchua_ garycmartin: do you have a list like that?
18:39 garycmartin mchua_: no but I've seen partial lists bounced about in email threads.
18:40 mchua_ welcome back again, tabitha_. :)
18:40 tabitha_ flakey connection today - so sorry
18:40 mchua_ no problem.
18:40 garycmartin mchua_: Based on diffing the file structure before and after boot.
18:40 mchua_ garycmartin, could you give me a pointer to where to find the threads?
18:40 er, those threads?
18:41 (devel, testing, etc?)
18:41 garycmartin mchua_: Sure, I'll go look and email you after.
18:41 mchua_ that's a good note to add to nandblaster stuff.
18:41 thanks!
18:41 anything else?
18:41 it's a small short meeting today, and there's... mostly, really, just a lot of work to be done
18:42 dirakx as allways :)
18:42 mchua_ tabitha_ has the list of Activity testing requests at http://wiki.laptop.org/go/Comm[…]#Testing_requests...
18:42 garycmartin mchua_: nandblaster idealy is used for pushing a clean image (say of a USB key), but deployments are excited about customising a live XO and pushing it out to all...
18:42 tabitha_ keeping an eye on xocamp - but if anyone wants to TELL us what to focus on in the coming weeks, we are happy to test whatever people want
18:43 mchua_ I have 8.2.1 (hurrah for bpepple's email with useful links), and throwing more scouts at Fedora testing...
18:43 garycmartin is pulling in the email threads for NANDblaster, poking Activity testing up towards the appropriate SL team, and...
18:43 ...I think that's it.
18:43 tabitha_ we will definitely keep testing food force, good to know gnubook and socialcalc on the list now
18:43 mchua_ Anything else?
18:43 yay tabitha_, thanks!
18:44 Oh - if anyone knows of any groups with access to network/wifi test/monitoring equipment who might want to handle a 50-XO testbed and can write up a plan on how they'll manage it / get community participation in it, try to get them to write up that proposal. :)
18:44 so we'll have something to take to ed.
18:44 otherwise, that's all I had.
18:45 Other business, questions, or the like?
18:45 garycmartin Hmmm, 50-XOs, that would keep my flat nice and warm ;-)
18:45 mchua_ Does anyone *not* have enough to do? ;)
18:45 marcopg that sounds unlikely ;)
18:45 tabitha_ we will take more XOs if they are available, but 50 is too much
18:45 we have about ten and could manage another ten easily
18:46 mchua_ aye, okay, so I guess 1cc stays as "The Big Testbed" for the time being, which is ok - until after we get 8.2.1 out the door.
18:46 okeydokey.
18:46 garycmartin mchua_: any one know what the mean deployment class size is?
18:46 mchua_ garycmartin: well, officially, in the 10k-ish range, I think
18:47 garycmartin: that drops quite a bit if you count tiny deployments with g1g1 machines
18:47 I don't know how much
18:47 oh! class size, you asked. not deployment size.
18:47 garycmartin mchua_: 10k, wow that's a large class ;-b
18:47 marcopg a bit busy :P
18:48 mchua_ Hm, I don't know exact numbers; can ask reuben, but from what I remember the deviation's pretty wide
18:48 marcopg garycmartin: why do you wonder about class size?
18:49 garycmartin mchua_: no problem. I'm guessing most teachers would break a class into working groups any way.
18:49 mchua_ garycmartin, I'll shoot reuben an email and cc the testing list, in case he has a quick answer.
18:49 tabitha_ I have a question about gnubook - should there be a .xo file somewhere?
18:49 mchua_ marcopg: I'd guess because that's the size of the testbed we'd want to test for
18:50 garycmartin marcopg: Shared activity testing (and use cases), what good for teachers re: N kids working on a project together.
18:50 mchua_ tabitha_: it's an ajax bookreader, so you open Browse and go to http://openlibrary.org/olpc/bookreader
18:50 marcopg mchua_: ah ok, than I guess deviation is not that important, as long as testbed covers the worst case...
18:50 mchua_ tabitha_: (it took me a moment to figure that one out)
18:50 marcopg: yep
18:50 garycmartin, marcopg, tabitha_ : officially, we test at 50
18:50 as a max
18:50 regardless of what size deployment classrooms are
18:50 garycmartin marcopg: Thinking about Labyrinth sharing use case, but same for activities we test that have sharing.
18:50 tabitha_ we looked at this a while back and wondered why we are looking at browser books
18:51 marcopg mchua_: which without having any clue about deployment seems pretty good
18:51 tabitha_ just checking nothing has changed with gnubook test request
18:51 mchua_ garycmartin, marcopg, tabitha_ : 10 in a simple mesh, 20 on an AP, 50 with an XS.
18:51 tabitha_: cool. :)
18:52 garycmartin: not sure if that helps, since I know a lot of folks don't even have 10 XOs (the welly testers collectively being a happy exception.)
18:52 marcopg the XS will have more than one class connected though?
18:52 I guess that's what confused me about looking at class size...
18:52 mchua_ marcopg: there aren't really clear test cases for that bit yet, really
18:52 (which, yeah, is a problem)
18:52 marcopg well the software sucks anyway, so not much to test ;)
18:53 you have a bit of time to come up with that :P
18:53 mchua_ looks around to see if martin's here
18:53 marcopg (hopefully better for 0.82)
18:53 0.84
18:53 tabitha_ just about to say, cant you ask martin what he looks at for XS load?
18:54 mchua_ aye, tabitha_ - that's something I don't have the bandwidth to do right now, though
18:54 though if someone else would like to pick it up, that's great ;)
18:54 tabitha_ he does not appear to be online
18:54 mchua_ generally speaking, if it's not 8.2.1 or moving stuff upstream, it's on my Defer To Later list (where later == think about again starting in Feb)
18:55 tabitha_ February which year?
18:55 mchua_ since those two should be more settled out by then, allowing us to do things like XS testing and looking at what to do with testbeds.
18:55 2009... I hope. ;)
18:55 marcopg hehe
18:55 mchua_ Certainly by 2020.
18:55 marcopg sounds like something post Feb to me
18:55 garycmartin mchua_: thanks, yea those numbers seem to make sense. It's more use case rather than technical max that's the sweet spot. Need to make sure that number gets plenty of testing (very sparse so far).
18:55 mchua_ sweet.
18:55 marcopg mchua_: don't overpromise ;)
18:56 mchua_ marcopg: underdeliver and overpromise, was it? ;)
18:56 marcopg hah
18:56 mchua_ anyhoo, I have to run, there's a lot of 8.2.1 testing and trying to book plane tickets for me tonight...
18:56 garycmartin mchua_: by 2020 I hope well be developing on you open source HW platform ;-)
18:56 mchua_ anything else you folks need / want from me besides 8.2.1 test results and minutes and action items from this meeting? ;)
18:56 marcopg would like to fly to nz instead of europe!
18:56 tabitha_ thanks for another great meeting Mel
18:57 All welcome in NZ
18:57 marcopg tabitha_: aaah, will be coming soon then ;)
18:57 mchua_ garycmartin: oh, if I get my PhD in what I think I'm going to, and you find yourself interested in hearing aids, that might be a possibility. ;)
18:57 marcopg martin made me want to go
18:57 mchua_ all righty, so that's a wrap in 8...
18:57 ...5....
18:57 ...3...
18:57 ...2....
18:57 ...1....
18:57 tabitha_ Send itineraries this way and I will help arrange tourist part of your visit :-)
18:57 mchua_ ....1....
18:57 *gavel!*
18:57 #endmeeting

Index | Today     Channels | Search | Join

Powered by ilbot/Modified.