Transcripts for What's Next? Session

From OLPC
Revision as of 14:27, 24 October 2010 by NinaStawski (talk | contribs)
Jump to: navigation, search
NinaStawski: we will discuss the past and future of the laptop and the development for XO 
[10:12am] NinaStawski: Daniel Drake introduces himself
[10:13am] NinaStawski: he helped make deployments in 7 countries or so — that's very impressing!
[10:15am] CanoeBerry: pleia2: thanks so much for helping! SJ too please when we find himm!
[10:15am] NinaStawski: the sustainability topic
[10:16am] andi_g: Daniel explains his work on the platform to make software updates smoother
[10:16am] NinaStawski: Daniel works at the latest version of Sugar, it will be based on Fedora 14
[10:16am] culseg left the chat room. (Remote host closed the connection)
[10:16am] NinaStawski: it's the very early stage of development for this version
[10:16am] NinaStawski: (I suggest it's v. 0.90—does someone knows?)
[10:17am] culseg joined the chat room.
[10:17am] pleia2: there are approx 80 patches to the linux kernel required for the XO, Daniel is doing work pushing them upstream to get them into the kernel
[10:18am] bensheldon joined the chat room.
[10:18am] NinaStawski: the focus of Daniel's work is to improve sustainability, not the release itself
[10:18am] NinaStawski: Steven introduses himself, he works on Fedora
[10:21am] NinaStawski: I think we need to get that presentation to post it on wiki)
[10:21am] NinaStawski: Burnie shows a presentation about Dextrose
[10:22am] NinaStawski: the link to the wiki page of Dextrose is http://wiki.sugarlabs.org/go/Dextrose
[10:22am] NinaStawski: the current version is much more stable than the previous ones
[10:23am] NinaStawski: he describes the recent addition - an accessibility control panel and it's improvements
[10:24am] NinaStawski: it was developed entirely in Uruguay
[10:24am] CanoeBerry: Bernie 
[10:24am] NinaStawski: oh sorry))
[10:24am] NinaStawski: I said my En is bad))
[10:25am] NinaStawski: on-screen keyboard
[10:25am] NinaStawski: a lot of the laptops are broken at places, and with this keyboard it's still possible to do it
[10:26am] NinaStawski: oh
[10:26am] NinaStawski: to work with it
[10:26am] sj2 joined the chat room.
[10:27am] sj2 left the chat room.
[10:27am] pleia2: picture of the virtual keyboard slide: http://www.flickr.com/photos/pleia2/5110741567/
[10:27am] NinaStawski: people are using it — we know it cause people are reporting bugs 
[10:27am] sj2 joined the chat room.
[10:28am] NinaStawski: bug reports is the feature that was also added recently
[10:29am] sj2: daniel and bernie talk very quickly and with animation 
[10:29am] NinaStawski: ..journal sorting options — sorting by size
[10:29am] sj2: <hey, irc.cgi -- a smiley face does not mean 'happy'>
[10:29am] sj2: <gets used to using forum.laptop.org/chat>
[10:29am] NinaStawski: it's useful to show kids a concept of size, cause they don't know whtat's big and what's small
[10:30am] sj2: there's a discussion of whether to show a logarithmic scale of size
[10:30am] NinaStawski: ...ad-hoc networks
[10:31am] sj2: ad-hoc wifi networks : channels 1/6/11 get three different icons, finally
[10:31am] sj2: so you can differentiate between them
[10:31am] agordon joined the chat room.
[10:31am] morganya joined the chat room.
[10:31am] NinaStawski: Bernie explains how ad-hoc network works
[10:32am] sj2: bernie explains that for ad-hoc networking, as with mesh before it,
[10:33am] sj2: in a small group you wont always know that something bbroadcast from one machine will be heard by another
[10:33am] sj2: dsd: inb the mesh,
[10:34am] sj2: 10 times a second every machine broadcasts to say "i am here" to syncronize the clocks
[10:34am] sj2: and help them communicate well with one another.  this is good -- clock synchronization is difficult in ad hoc networks, to ketp the order of message s 9for instance) steaight
[10:34am] sj2 left the chat room. (Quit: CGI:IRC)
[10:35am] sj2 joined the chat room.
[10:36am] sj2: in ad-hoc, one machine is generally the 'king' of the network.  and there are listeners; if someone istening doesnt hearsomething, he become the 'king' and starts broadcasting
[10:36am] sj2: thi is something that does not work so well for adhoc.  performance can be much worse than mesh for arge groups
[10:36am] sj2: in that you can have network splits, and you may find it less reliable for the mesh
[10:36am] sj2: but a good substitute when you have 2 kids or 3 kids under a tree.
[10:37am] NinaStawski: but for a small amount of laptops it should be the right solution
[10:37am] sj2: there is also a problem unresolved so far that ad-hoc is slow (at scale?) on the xo... the wireless card we have is still a bit of a black box, which akes it hard to diagnose.
[10:38am] atphalix joined the chat room.
[10:38am] andi_g left the chat room. ("parting")
[10:39am] sj2: weve stop making the mesh run inside the kernel -- we are making xos sleep a lot, ev en when you dont know it (between key presses). 
[10:40am] sj2: and we now need to expose a very new mesh interface to the ser, which will take soe time before it is as smooth as what benie is showing for an ad-hoc interface of 3 channels
[10:40am] sj2: user*
[10:40am] sj2: adam - now that you have full 3g modem support, do you want one computer that has a 3g modem to be able to share that connection over adhoc or mesh?
[10:40am] sj2: bernie -- martin a. told me last week that he developed this while bored at a conference.
[10:40am] sj2: as it is aid 'we are all standing on the shoulders of giants' in the fedora and linux community, so martin was able to add this advanced feature in no tie.
[10:41am] sj2: one xo connects to 3G and shares this with others.
[10:41am] sj2: dsd - you would have to hack something on the client side... to amke the client look for dhcp.  it doesnt do that automatically.  i dont think there is anythings olid, we are cose to making that possible,
[10:41am] sj2: but htis is the fist time i heard about this in a while.
[10:41am] sj2: bernie - we wer doing this in paraguay, but we were connecting the xo to the access point.  youre right.
[10:41am] sj2: dsd  there are a couple of small tech points we have to solve first, but nothing big.
[10:42am] sj2: bernie - we have a backup/restore feature as well.  how does it work?
[10:42am] NinaStawski: ..backup and restore to usb sticks
[10:42am] NinaStawski: it's two clicks backup
[10:42am] sj2: we have one-click or two-click backup now -- from the journal, you have a hover menu
[10:42am] sj2: that lets you backup to a server.  then says
[10:42am] sj2: 'please wait..' while backing up.  this is manual backup on demand. 
[10:43am] NinaStawski: the backup is usually automatic, but it does not work well
[10:43am] sj2: dsd - some schools also do backups automatically once a day
[10:43am] sj2: the manual 'Restore from backup' here is a key difference
[10:43am] NinaStawski: so adding the manual backup was essential
[10:43am] NinaStawski: there is also a restore option from the same menu
[10:44am] tch joined the chat room.
[10:44am] NinaStawski: q: when you are restoring, what does it do with the new files?
[10:44am] sj2: <side debate about why the icon looks like a school with a belltower or a church... what, no minaret options?>
[10:45am] sj2: bernie: there may be data loss - if a user clicks randomly they can lose files.  backuprestore opens a possibility of losing data this way
[10:46am] NinaStawski: the journal is deleted and fully restored from backup while restoring
[10:46am] sj2: tincha wantd to be able to do partial backup restore.
[10:46am] NinaStawski: there are no partial backups yet
[10:46am] sj2: this change hasn been accepted upstream yet for some of thee reasons. it doesnt do a timestamp check to ensure thqt yourr journal doesnt have new items since the b ackup was made...
[10:46am] sj2: but it is better than the 'nothing' we had before. 
[10:47am] sj2: (this is all bernie talking)
[10:47am] NinaStawski: laptops already perform backups every day, so you don't loose a big amount of data anyway
[10:47am] sj2: bernie - this works for a specific depoyment now with its own regular backup context [its not ready to be used everywhere]
[10:47am] sj2: pablo - when you backup on a scholserver how does it identtify your laptop? 
[10:48am] NinaStawski: laptops are authenticated at the school server by a serial number
[10:48am] sj2: bernie - there is a registration process.  but many laptops dont backup now because some children dont register in the start of qa class
[10:48am] sj2: dsd 0 we work with 6 yr lds etc.  registration does not make any sense it is not clear why it is necessary - and to be honest we should
[10:48am] sj2: have it automated.  and when you click it it does nothing obvious in the ui, so what was the point in that?
[10:48am] NinaStawski: there is no visible benifites for users, so they just don't do it
[10:48am] grantbow joined the chat room.
[10:49am] sj2: (there is a process to register of your laptop with your school server, to give you an identifying number)
[10:49am] sj2: bernie - we want to make the schoolserver part of the general ui. 
[10:49am] NinaStawski: to make it visible, the idea is to make their icon visible in the network after registration
[10:49am] sj2: anna s -- is there goig to be an 'unregister' option?  somekids pass between two different schoolservers
[10:49am] sj2: bernie 0 after reg, rather than removing the button, we will change the label to 'register again'.  a bit ugly
[10:50am] sj2: grant - but you have avbackups on the old schoolserver then... whats the modality for getting those backups yet?
[10:50am] sj2: bernie 0 schoolserver is called version '0.6' for a reason.
[10:50am] sj2: new children keep coming and oldones go away and you have backups forever...
[10:50am] morganya: btw, the "Martin Abente" that Bernie is talking about just logged in as tch 
[10:50am] sj2: grant - you could also have a usb stick option so kids can take it with them.
[10:50am] sj2: bernie  yes we now have 'backup to USB' which allows this
[10:51am] sj2: <waves at tch !>
[10:51am] tch: hello amigos!
[10:51am] sj2: bernie - a lot of children learn the shell quickly when they want to do something with it.
[10:51am] NinaStawski: another ugly fuction - repair the journal 
[10:52am] NinaStawski: sometimes it gets broken because of the hardware problems
[10:52am] sj2: pablo - dometimes old datastores get cached and a new one started... does this delete the old ones?
[10:53am] sj2: bernie, dsd - no. 
[10:53am] sj2: grant - that bug can ake you run out of space... which amplifies the problem
[10:53am] sj2: bernie - one problem is the Wine activity.  you remove wine, bc the teachers see it cluttering the laptop with stuff...
[10:53am] sj2: but that only saved 10M because the kids were storing the windows apps elsewhere.
[10:53am] sj2: now we remove those cached files as well...
[10:54am] NinaStawski: now we remove temporary files - before there was a problem when you delete browser, and the cache is still there
[10:54am] sj2: [note to self -- wine is cool, we should not cripple it... maybe inmprove the way it gets uninstalled.]
[10:54am] sj2: bernie - we want to remove all the ancillary stuff with an activity when the activity is removed
[10:55am] sj2: bernie - there are problems with backup restore  upstream devs wanted something done as an activity...
[10:55am] sj2: where you can backup individual files.
[10:55am] NinaStawski: ..intuitive resource monitoring
[10:56am] sj2: sj -- noting with bbarry that this sort of backup restore is important.  so we should clarify what the different streams of conversation are
[10:56am] sj2: and what
[10:56am] sj2: upstream
[10:57am] NinaStawski: bernie: it really is possible for deployments to customize Sugar
[10:57am] sj2: 'upstream' means in such a small community
[10:57am] sj2: <bernie gives an example of deployments all of which had lots of customizations already>
[10:57am] sj2: dextrose is a merge of many distributions and customizations that pleases at least 3 deployments... saving time in testing.
[10:58am] NinaStawski: resource monitoring-now it looks to geeky
[10:58am] sj2: 'system resources'
[10:58am] sj2: helps you see cpu in use and memory in use.  it has not been accepted upstream yet because it is too geeky?
[10:59am] NinaStawski: there are three icons showing the status of the laptop
[10:59am] NinaStawski: depending on the CPU/Memory load
[11:01am] sj2: question about asking users which activity they want to close when opening a new one
[11:01am] NinaStawski: suggestion: it's maybe better to not open the activity if the memory is already full
[11:01am] sj2: benieL we are talking about [Out Of Memory] alternatives -- whether to start with the oldest or the latest activity
[11:02am] NinaStawski: Finger / Stylus mode for touchpad
[11:02am] sj2: sj - note that this automatic assumption that "an activity should be killed" is a bit murderous -- the original question was about when to ask users if they /want/ to cose an activity
[11:03am] sj2: reinding the that they are running out of space, rather than choosing an activity to kill when things get relly bad
[11:03am] bensheldon left the chat room. (Quit: Colloquy for iPhone - http://colloquy.mobi)
[11:05am] NinaStawski: Improvements to the GSM support and also GSM sharing
[11:06am] NinaStawski: also the tabbed browsing was added
[11:06am] NinaStawski: everybody likes it except UI designers)
[11:06am] sj2: bernie has another patch that allows you to switch between relative-capacative and relative-stylus modes (in both cases just moving the ouse around with a finger or stylus, not 'draing' in absolute position mode)
[11:07am] NinaStawski: (some pictures from Pagaguay)
[11:07am] NinaStawski: Paraguay
[11:07am] sj2: there is also the option of a 'drawing' mode, which we used to have as an option... but it was sometines jumpy, and that ability was removed at some point from the kernel.
[11:07am] sj2: sj - why not offer all three options, two relative modes and one absolute mode? 
[11:08am] sj2: bernie - we didnt think about that, a good idea.  dsd - it is in flux what hardware modes are supported; I subnitted a kernel patch that only had relative modes and they did not accept it
[11:08am] sj2: so the absolute mode may be possible again in the future
[11:08am] sj2: <paraphrased>
[11:11am] alsroot joined the chat room.
[11:12am] walterbender joined the chat room.
[11:12am] walterbender: ciao
[11:12am] sj2: hi walter
[11:13am] sj2: wlcome   bernie is talking about the linux foundation... in a gap in the presentation
[11:13am] sj2: welcome back als
[11:13am] alsroot: sj2: hi
[11:13am] anna_bham left the chat room. (Ping timeout: 255 seconds)
[11:13am] sj2: laptop with stage fright.... the session is almost over, so here is a quick rundown from my notes
[11:13am] sj2: 150 years ago today the first transatlantic cable was sent
[11:14am] NinaStawski: James started his talk
[11:14am] sj2: and the message itself was lost... 
[11:14am] Cerlyn joined the chat room.
[11:14am] sj2: since then we have gotten away from asynchronous operation like tat tickertape machine... though we use similar things on irc
[11:16am] NinaStawski: Teaching computers to teach people to read and speak
[11:17am] NinaStawski: speech recognition and pronunciation evaluation
[11:18am] NinaStawski: Julius open source speech recognition
[11:18am] NinaStawski: triphones vs diphones
[11:19am] NinaStawski: James askes us to support the Gnash, as there are no real ways to get microphone upload other than with Adobe Flash

URL: [[1]]