Slashdot Mirror
TechCrunch Wants To Create an Open Source Tablet
RKo618 writes "TechCrunch announced that they are planning to design their own $200 web tablet device. Quoting: 'The idea is to turn it on, bypass any desktop interface, and go directly to Firefox running in a modified Kiosk mode that effectively turns the browser into the operating system for the device. Add Gears for offline syncing of Google docs, email, etc., and Skype for communication and you have a machine that will be almost as useful as a desktop but cheaper and more portable than any laptop or tablet PC.' The aim is for the tablet to run on modified open source software, which will be released back to the community along with the specifications for the hardware."
They have to compete with the N770 and N800 both that run open source software and both already have a very large installed base of users.
They have to compete with that, so they really need to get it right. I love my N770 except for battery life. I wish these things could go at least 3 days between charges.
Do not look at laser with remaining good eye.
Firefox is a great operating system it lacks a good browser, though.
Asterix for voice communications.
I hear Obelix is way better.
(But other than that I agree. Skype is not open source and a security liability.)
Except the open source bit?
An application you really, really do not want to debug.
... is that it's not possible.
Look at the specs - if you want decent battery life, a decent screen (with decent resolution), decent RAM and storage (specced at 512MB and 4GB), and to all go for $200, it's hard.
The only thing on the market NOW that's even remotely close is the XO-1, but it only has 256MB of RAM and 1GB storage. And it's BOM costs are quite high already, even with its anemic CPU. If you want to mass-produce it and sell for $200 retail, after taking out everyone's profit and overhead, you're looking at a manufactured unit cost of around $100. Maybe $125, if you can squeeze profit margins from retailers and the like. (Figure in profit/time for doing the software, as asll as distribution costs to get it to retailers - you'll probably want wholesale to cost around $150-160). Of that, the screen, RAM and flash are the big budget items, and a good CPU can be pricey in quantity ($10-ish, nominally for a high-end ARM processor from the big companies - Samsung/Marvell/Freescale).
It's a tight squeeze, add in the other costs like warranty and support, and you'll find not many people are willing ot take on such a high-risk project with such little returns. You can try to sell it online like the OLPC guys with their "give one get one" thing, which lets you raise the manufactured cost more, but then have to deal with all the issues of distribution to end users.
It's not that no one wants to do it, it's just that it's really hard to do a good job in very tight constraints. Give it a year, and you'll probably be able to do it with last year's CPUs, last year's RAM, and last year's storage. But if you up the requirements next year, well.
The original Eee PC had a crappy screen, crappy battery life, OK CPU, as-required RAM and as-required storage, and still cost $400, even though the screen was bulk leftovers from portable DVD players, and the CPU was more or less "hey, I found a box of these things sitting on the shelf".