Grassroots unconference/sunday notes

From OLPC
Jump to navigation Jump to search

also see http://wiki.laptop.org/go/User:Sethwoodworth/june8

Jun 08 11:04:53 *	mchua is transcribing from grassroots unconf again
Jun 08 11:04:56 <mchua>	cliff: ...on the grassroots side, talking about Ghana and how we go into groups in Ghana
Jun 08 11:05:02 <mchua>	content is the mots important thing
Jun 08 11:05:12 <mchua>	going to talk with local people about what the problem are there
Jun 08 11:05:18 <mchua>	things like "all the men are getting X, but the children aren't"
Jun 08 11:05:27 <mchua>	what crop of soybeans will get the greatest yield things like that
Jun 08 11:05:35 <mchua>	what efforts... the contnet is being recorded on ohw to prevent the spread of hiv/aids
Jun 08 11:05:46 >isforinsects<	#olpc-groups
Jun 08 11:06:03 <mchua>	b/c the content is so important one of the ifrst things we did was to get press out
Jun 08 11:06:06 <mchua>	probably less relevant for olpc though
Jun 08 11:06:12 <mchua>	in april, I went back to ghana to the rural areas
Jun 08 11:06:20 <mchua>	b/c we didn't have our device ready yet and couldn't get them to record content
Jun 08 11:06:25 <mchua>	we just got some cheap voice recorders
Jun 08 11:06:33 <mchua>	(note: they are making a $5-$10 voice recorder)
Jun 08 11:06:43 <mchua>	just asked them to start recording anything they were speaking about
Jun 08 11:06:53 <mchua>	just to start recording what they were planning to tell people
Jun 08 11:07:00 <mchua>	so the way in which we went about hti swas to pick some districts
Jun 08 11:07:08 <mchua>	we looked to districts wiht low adult literacy rates
Jun 08 11:07:24 <mchua>	we wanted to work in places where the chances of using the internet or having some ict lab were low
Jun 08 11:07:28 <mchua>	not realistic anytime soon
Jun 08 11:07:34 <mchua>	so that's where we start is finding those districts
Jun 08 11:07:40 <mchua>	i went to the idstrict assembly of each to meet with the chief, etc
Jun 08 11:07:42 <mchua>	executives
Jun 08 11:07:50 <mchua>	and the nfrom there started talking to the groups that heard from the villages
Jun 08 11:07:59 <mchua>	so the information is typically health/small-biz/agriculture
Jun 08 11:08:01 <mchua>	these are the 3
Jun 08 11:08:10 <mchua>	obviously there's education b/c of the lit. use of the device.
Jun 08 11:09:22 <mchua>	*brings out prototype of audio recorder*
Jun 08 11:09:28 <mchua>	(large, orange, big buttons)
Jun 08 11:09:35 <mchua>	want to be able to let people navigate audio like they navigate a book
Jun 08 11:10:01 <mchua>	(3d printed!)
Jun 08 11:10:07 <mchua>	talk about how we want people to be able to navigate hte audio
Jun 08 11:12:50 <mchua>	issues: large # of regional languages
Jun 08 11:13:20 <mchua>	mel: how did you find villages?
Jun 08 11:13:27 <mchua>	cliff: had v. few conversations with the central govt.
Jun 08 11:13:33 <mchua>	worked with ministry of education, but we don't work from the top down
Jun 08 11:13:44 <mchua>	we use statistics that these orgs have done, unicef, some big group has done
Jun 08 11:14:01 <mchua>	I can see the stats, and now that I have a few contacts I can go to the facultyof a uni and they can tell me more
Jun 08 11:14:09 <mchua>	by looking at the #s I can see these are the districts are the ones most in need
Jun 08 11:14:11 <mchua>	this isn't really scaleable,
Jun 08 11:14:17 <mchua>	but what I've bene doing is that I just go into the villages
Jun 08 11:14:18 <mchua>	and talk to people
Jun 08 11:14:25 <mchua>	and get a sense for what info I need, who is serving them already
Jun 08 11:14:32 <mchua>	most of these villages - there are people serving them, giving tem info
Jun 08 11:14:37 <mchua>	but they're govt agencides, nonprofits...
Jun 08 11:14:56 <mchua>	Ithen i go to them, say "hey, we're looking at this distrcit b/c it has need fo rhtis, and you serve this district... would you like to participate?"
Jun 08 11:15:30 <mchua>	this isn't the long term thing
Jun 08 11:15:42 <mchua>	in the future we'll have the local groups do that, find the villages - and then come to us
Jun 08 11:15:47 <mchua>	mel: when and how are you going ot make that switch?
Jun 08 11:15:53 <mchua>	cliff: I think what it'll be is sthat ew're going to make...
Jun 08 11:16:03 <mchua>	oh! we're working with india now, ghana and india are different
Jun 08 11:16:14 <mchua>	they both have literacy issues, there are enough differences to make them not work wiht one model
Jun 08 11:16:30 <mchua>	("but there are enough differences...")
Jun 08 11:16:40 <mchua>	the structure might be that in separate nonprofit orgs starts up in that country
Jun 08 11:16:48 <mchua>	and that they have their own governance model and they deal with how the devices get imported
Jun 08 11:16:54 <mchua>	and they know how to reasearch their own districts
Jun 08 11:17:00 <mchua>	and so I'm doing a lot of hope that happens in ghana right now
Jun 08 11:17:07 <mchua>	we'll probably do that in one of the states of india and around bangalore
Jun 08 11:17:12 <mchua>	if that's successful wthe nwe 'll tyr to apply that same thing
Jun 08 11:17:19 <mchua>	and when people come to us and say hey, we've heard about you, how can we start
Jun 08 11:17:25 <mchua>	we say here's a case study of how th eorgs tstart in these countries
Jun 08 11:17:39 <mchua>	but i dn't know how this will work out
Jun 08 11:17:46 <mchua>	christoph: 18 months down the road, projections?
Jun 08 11:17:53 <mchua>	cliff: 12 mo donwn the road from now we'll contact mfgr.
Jun 08 11:18:06 <mchua>	if I wanted to, we could go to content manufacturing and produce 100k units
Jun 08 11:18:08 <mchua>	by the end of he year
Jun 08 11:18:16 <mchua>	but we're refiing the design here
Jun 08 11:18:21 <mchua>	what we pilot test isn't going to be final
Jun 08 11:18:40 <mchua>	seth: do you own the copyright or does another company own the content?
Jun 08 11:18:45 <mchua>	cliff: all us
Jun 08 11:18:48 <mchua>	seth: will be open source?
Jun 08 11:18:54 <mchua>	cliff: we are open sourcing everything we do, hardware software content
Jun 08 11:19:03 <mchua>	the hard thing about hardware... you get the price down, of course we're not designing our own chip
Jun 08 11:19:14 <mchua>	so we're using a chipset... a chipset a taiwanese mfgr makes
Jun 08 11:19:29 <mchua>	they have proprietary codes, but they're not royalty based so you can move things in/out of them
Jun 08 11:19:45 <mchua>	if we used free codecs, we'd have to raise he price of this by like $50
Jun 08 11:19:51 <mchua>	(b/c the chips just don't exist for low prices)
Jun 08 11:20:05 <mchua>	so we have a lot of c, a lot of assembly for this specific chipset
Jun 08 11:20:15 <mchua>	what we're trying to do to open this up is that...
Jun 08 11:20:26 <mchua>	even as I said you can use this as a literacy device, and naybody can say this is how I want to be
Jun 08 11:20:36 <mchua>	this button is a word left/right, or tihs button is a page left/right
Jun 08 11:20:44 <mchua>	we have software that will generate those kinds of configurations
Jun 08 11:21:20 <mchua>	so the metadata is pretty simple
Jun 08 11:21:30 <mchua>	w're writing an audio operating system for tihs in some language
Jun 08 11:21:41 <mchua>	adn when you turn his device on and say "welcome to the talkkng book" and tells you waht content you have on it
Jun 08 11:22:09 <mchua>	we have verbose and brief modes for when you want to turn guidance on/off, b/c talking uses battery
Jun 08 11:22:26 <mchua>	but we have a user0configurable xml file 
Jun 08 11:22:29 <mchua>	it's just a special case of user content
Jun 08 11:22:53 <mchua>	so we open up things by putting things in a higher level language
Jun 08 11:23:16 <mchua>	it's a special case of user content b/c it's just a recording of someone speaking about what button to press when
Jun 08 11:23:28 <mchua>	if you want an os that speaks to people in all their diff. languages you need to let people in those groups record it
Jun 08 11:23:39 <mchua>	get someone in that village who knows english to make the recording for their village
Jun 08 11:24:03 <mchua>	so when someone goes to a kiosk library to download content, they can also say "hey, change my os to this language too"
Jun 08 11:24:11 <mchua>	christoph: so the main point of content distribution is the kiosk?
Jun 08 11:24:22 <mchua>	cliff: they can talk device to device w/ usb
Jun 08 11:24:30 <mchua>	we leave it up to the local orgs to do the kiosk
Jun 08 11:25:03 <mchua>	christoph: would you ever enforce a policy that at the kiosk you have to take care of it, etc?
Jun 08 11:25:15 <mchua>	cliff: we're going ot propose a business plan for how pilots could work
Jun 08 11:25:20 <mchua>	<but we're not going to implement it>
Jun 08 11:25:37 <mchua>	so with that, we might say that a kiosk owner is like a franchise of this organization
Jun 08 11:25:47 <mchua>	and to get that franchise license ou need ot meet requirements X, Y, Z
Jun 08 11:25:52 <mchua>	so the content downloading is always free
Jun 08 11:26:13 <mchua>	(christoph's question was about whether you might run into people holding content kiosks hostage by charging for their urse)
Jun 08 11:26:49 <mchua>	in rural villages they use cellphones to transfer data from kiosk to kiosk
Jun 08 11:27:13 <mchua>	that costs a lot
Jun 08 11:27:21 <mchua>	therefore, not sure how well that model will work out
Jun 08 11:27:38 <mchua>	christoph: what format are the files in?
Jun 08 11:27:46 <mchua>	cliff: propietary, since that's what the codecs on the chipset come on
Jun 08 11:27:56 <mchua>	you can convert them to wav though, that can be done ,no legal restrictions
Jun 08 11:28:29 <mchua>	we can put that conversion software on every device, so you can always have modifiable files with you
Jun 08 11:28:42 <mchua>	right now thinking sd cards for storage
Jun 08 11:28:45 <mchua>	that's minimum 1 gig, that's huge
Jun 08 11:28:48 <mchua>	that's like 100 hours
Jun 08 11:29:07 <mchua>	what I really wish - the biggest tech problem is instead of getting more memory at the lowest price
Jun 08 11:29:35 <mchua>	I wish there was some way you could take $1.50 of that cost off but get much less memory
Jun 08 11:29:44 <mchua>	lower amt of memory w/ lower price
Jun 08 11:30:15 <mchua>	looking for cheap, non-upgradeable flash memory - tiny size, tiny price - 3 hours, 10 hours of memory
Jun 08 11:31:11 <mchua>	then you can upgrade on your own
Jun 08 11:32:06 <mchua>	maybe we can have a grassroots network of people relaying microsd cards between kiosks, physically
Jun 08 11:47:55 <mchua>	<discussion now on hackability of the device - will transcribe later>
Jun 08 11:55:14 <mchua>	hackability discussion recap:
Jun 08 11:55:31 <mchua>	mel: you're encouraging high level hacking, how about low level hacking, can people get involved in that?
Jun 08 11:55:38 <mchua>	cliff: yeah, we have all our code which will be open even the assembly etc.
Jun 08 11:55:58 <mchua>	mel: right, but there's a difference between open code, and support for people who want to work on open code - documentation, a way for them to ask questions?
Jun 08 11:56:14 <mchua>	cliff: oh we will be doing that, we want to have a meritocratic low-level dev community
Jun 08 11:56:22 <mchua>	seth: plan to let hackers buy prototypes so they can develop on it?
Jun 08 11:56:32 <mchua>	cliff: it would be $150 per unit or something right now
Jun 08 11:56:38 <mchua>	mel, seth: look at g1g1; they will pay
Jun 08 11:56:49 <mchua>	cliff: ooh.
Jun 08 11:58:21 <mchua>	mel: this is dangerous, though, because you get hackers concentrating on making their stuff work for them, instead of for kids in africa
Jun 08 11:58:26 <mchua>	sj: how is that bad?
Jun 08 11:58:49 <mchua>	mel: it's a distraction - see g1g1, you get people clamoring for webkinz support, olpc employees are spending time dealing with those instead of helping thailand get connectivity, for instance
Jun 08 11:59:21 <mchua>	it's not that the work they're doing is bad, it's that you need to make it clear that they're not your core mission to support, that if you want to work on things that wil make the experiecne better for people otehr than these folks in afraica you're actually trying to serve, that you'r eon your own
Jun 08 11:59:41 <mchua>	seth: high profile hacker bloggers like bre pettis, ladyada, hackaday, evil mad scientist labs, etc. can help