One Laptop Per Child (OLPC) has been showing off the first models of its XO 3 tablet at the Consumers Electronics Show in Las Vegas. The 8-inch tablet uses the 1GHz Marvell ARMADA PXA618 processor, 512MB of RAM and comes with a 1024x768-resolution LCD, or a lower cost Pixel Qi sunlight-readable display. Connectivity is via an …
If the OLPC is indeed still aimed at disadvantaged children in developing countries, then I strenuously object to turning them into very tiny Google cash cows by giving them Android (that the amount of cash in question is astronomically small, is beside the point).
Give them that other flavor of linux and thus the proper respect.
Why does Google get anything?
You don't have to pay Google a cent to use Android on a device. It might be different if OLPC were to certify the device, licence Marketplace and the Google Apps but I somehow that is a relevant concern for a device for kids.
More likely the device will have a fixed set of apps with no ability to modify the list, or if there is a "store" it will be a custom job that points to a teacher / administrator controlled set of apps that students may install on their machines.
Google gets 100% of the data
Obviously I wasn't referring to any up-front frees. Android is designed to make you use Google services - not to coerce anyone, true, but the entire OS is geared towards this purpose, and that is the most likely practical outcome. All Google services gather information on its users, which they then make money from through advertising, deals with third party advertisers, etc. (I can't believe that I actually have to explain Google's business plan to someone today).
So the statement "an Android user is a tiny cash cow" is factually correct, it just requires looking at the whole picture.
Apparently we have two options for the OLPC: an OS that turns dirt-poor kids into advertising objects, and one that is equally functional but doesn't. I know which one is the respectful choice.
And exactly what point would there be in advertising to "dirt-poor kids"?
Still explaining the finer points of the Google business plan, then. I don't just mean the targeted ads shown on the side of Google services (which is no doubt what you're referring to) but that the behavior of the user and all his data is analyzed. From this analysis, conclusions can be drawn that have advertising value (not necessarily to Google itself but possibly to other players in the advertising industry).
For example, from the OLPCs dropped into an Ugandan village, information can be extracted about how Ugandans think, what they'd like to buy, etc. This information is valuable even in the context of a developing country. True, given the very low spending power of the Ugandan consumer, the value won't be very great (which is the reason why I originally said "the cash will be astronomically small".) But it is being extracted, like crude oil from Nigerian soil.
Seriously, try to think this through a little and not just leap for the obvious.
I think the point is that Android is in one sense built around the idea of datamining the individual and profiling them in such a way that Google can use the information to sell ads.
Whether or not there's another tablet-oriented version of Linux that is as easy to support (from a hardware manufacturer perspective) and with as good a selection of software for the goals pursued by the OLPC project is for someone other than me to say...
Google does not get 100% of the data
Google only gets the data if you bother to use Google services from your Android build. e.g. if the browser's home page is set to google.com. Or if you install their apps which implies certification.
Lots of devices running Android have been configured to not use Google services. e.g. older Archos devices have their own app store and homepage. Some devices even use Android builds that talk with rival services (e.g. Baidu or Amazon).
I assume any OLPC build would not be using Google's services given the purpose of the project and the sorts of locations / environments they would end up being used from.
The datamining comes from using the value added apps which are not part of the Android open source code - Marketplace, GMail, Google+, Maps etc. i.e. you get driving directions from their maps app and they learn a little more about you and your location with which to sell advertising. There might be a limited number of data mining opportunities in the vanilla Android build (e.g. the default browser homepage) but there really isn't much else to speak of as far as I'm aware.
OLPC would probably take the default android source and build and modify that to their requirements. Google likely wouldn't earn much if anything from the effort beyond delivering ads to web pages which students happened to be browsing.
Which is why I said...
"Android is designed to make you use Google services - not to coerce anyone, true, but the entire OS is geared towards this purpose, and that is the most likely practical outcome."
Not every Android user will use these services. But it is the obvious choice and most will take it, because the OS is designed for this purpose. There would have been no benefit to Google if it hadn't been.
And personally I would guess that the people designing the OLPC have no great incentive nor much ability to change the default Android install. They will add their own learning-oriented applications to it, sure, but a whole app store? Unlikely. Everyone needs basic applications like e.g. email, so will the OLPC folks go out of their way to include a very visible alternative to Gmail? Probably not, because there is no special learning purpose involved in email (apart from "being able to use email", clearly, but that's just one lesson). More likely the kids will end up using Gmail (assuming the village has at least one wifi access point, obviously, but if it doesn't then the whole project is severely handicapped anyway, regardless of whether Android is being used.)
I would say Android has a useless collection of software for tablets never mind free and non adware good educational software for tablets.
The point is OLPC wouldn't get the Google apps at all. No GMail. They'd have the bog standard POP3 Mail app but not GMail, nor any of the other apps which do the datamining.
I doubt they will make much effort to customize the default build but it is easy enough to cut out apps which are surplus to user requirements, maybe write their own IME devices. There are enough alternatives to Marketplace to show it is not hard to write *something*. Even a basic web site with some .apk links would do the job but a thin app could be written that invokes the PackageInstaller activity for the same effect.
I have no idea how OLPC might port their UI to Android - Sugar is such a godawful mess that it would almost require a full rewrite. Maybe they haven't thought about it too deeply themselves.
and where are their fanbois saying that children need to be learning the Windows user interface and software because that's what they'll be using once they get out of school? What? Windows is moving to yet another user interface and software application framework? No wonder they're keeping their mouths shut this time.
If only Negroponte listened to all the people(inside OLPC and outside) who said that Microsoft on the XO was nothing but a stalling technique and they never intended to produce a viable product.
The tablet looks great and the video showing the Sugar Labs Sugar interface and applications looks great. They also put lots of effort into design again with how the connectors can or can't be covered by the protective cover and how an external battery can recharge the device.
Negroponte lost my respect four years ago
I was an active supporter of OLPC until it became clear that education wasn't Nicholas' primary objective for the project. I fear that Ivan Krstić's "Sic Transit Gloria Laptopi" blog entry is as relevant as ever.
Innovation in netbooks since they came out has been extremely limited. Why can't we see something like this in shops everywhere? Although I wonder how it comapres to the PI in temrs of cost/gunt/power useage.
Said it last time, I'll say it this time
OLPC really should be selling consumer versions of this tablet. I assume it would run Android reasonably well and its ruggedised form factor and cheap price would be attractive too. They'd have to charge more for a consumer version of course and tweak the specs / appearance a bit to keep the educational / consumer versions distinct (i.e. to stop people stealing educational models to sell on EBay). But I bet they could sell a lot of devices if they priced them for $200.
Sales would help them increase volumes and profits would subsidize the educational version so everyone would win.
What they should not do is launch some lame Buy Two Get One scheme like they did with the XO-1. OLPC pioneered the netbook form factor and watched impotently from the sidelines as others such as ASUS, Acer, HP et al profited from consumer demand for those devices.
Perhaps I missed something, but if the area where these are targeted is so remote you can only get there by helicopter, and the infrastructure is so basic you have to power the laptop from solar cells, how do they 'learn to use the Internet'? Satellite modem???
I don't think you have missed anything
because it certainly isn't clear from the article how it will work.
However Sugata Mitra has a track record of this kind of thing so you can be pretty sure that it isn't as silly as El reg has made it seem, take a look at http://www.ted.com/speakers/sugata_mitra.html
In short, mesh networking. Requires a lot of (active) devices to maintain a semi-direct link to wherever the nearest ISP is, however.
Assuming there will eventually be several hundred of these devices in a given area, and that they exert the same infectious effects upon the starving as it does to our youngsters.. this would not be a problem.
Next question on the list: If one offers the device in a trade, how much rice/wheat will be offered in return?