i've done a lot of investigation and thinking, and also been asking questions, and have come to the conclusion that there isn't an ARM or a MIPS or any other kind of low-cost non-x86 laptop which any free software user would be happy to buy... and have also come to the conclusion that there won't be one, either, without taking some sort of action that bypasses the current flawed development, production and delivery chain.
the problem is this: retailers simply won't risk selling something that doesn't run windows, and that's the end of it. they haven't got the money to take the risk on something as "unproven" as an ARM or MIPS laptop, especially when the costs of such laptops are so low and the margins are so ridiculous [10% for everybody. not 10% for the manufacturer, 10% for the agent, 10% for the retailer: 10% for _everybody_] that the only way the retailers can make money is by mis-selling you massively on a 2 or 3 year warranty scam.
then, the cost of designing a laptop from scratch includes costs like $100,000 for the plastic moulding (it's an art not a science). thus, the manufacturers have to sell at least 40,000 units at $120 each with a mere 5% markup in order to get a return on investment.
in short: if you want an ARM or a MIPS netbook from a retailer, at a reasonable price and with the kinds of features (decent 1280x768 or better screen; 8+ hour battery life) that you'd expect of an x86-based laptop, you can flat-out forget it.
thus, i come to the conclusion that the only way to get one is to get it made - just like the open-pandora custom-made a stonking games console for 4,000 people, which is just going into mass- production, now: sta tus.
now, as there would need to be consensus on what people would like, and it's necessary for other people to also know what everyone _else_ would like, it pains me to have to invite people to contact me with either a simple aol-style "metoo" posting, or an example specification and/or features that they'd prefer, along with a maximum price that they'd be willing to pay. please bear in mind that you should _only_ do this if you _actually_ would part with the money - possibly up-front (just like the open-pandora) in order to help fund the development / purchasing.
just so you know: i've already done the same thing on the open-pandora forum, and the response was actually very positive. even the troll with his tiresome campaign to have open-pandora discredited as a ponzi scheme was surprisingly useful, as people came out in defense of the pandora's creators and the method by which the money was raised. 4000 people trusted the open-pandora creators enough to pay up-front. isn't that something?
the Open Pandora forum posters basically came up with something like this, which i kinda could have predicted, but wanted _them_ - as you - to "have a say":
- * 1gb of DDR2 RAM
- * minimum of 1200 x 768 screen (i.e anything that calls itself WXGA)
- * matte finish LCD _not_ glossy.
- * 10in+ screen size (but keeping to a size which doesn't eat power. 10in = 4-5 watts; 11in = 5-6 watts...)
- * 8+ hour battery life (see above for biggest power hog)
- * 3D graphics capability
- * decent keyboard (e.g. the one on the Asus 1000HE)
- * nipple as well as trackpad - and a _decent_ trackpad
- * option for an internal SATA drive
- * at least 2 full-power-spec'd USB2 ports (close together)
- * option for at least 2 SD/MMC card slots (at least one for booting)
- * option for internal 3G, space for internal WIFI, option for etc. etc.
- * HDMI out
now, pricing all that up i come to _approximately_ a $212 BOM if you use a 1ghz Samsung S5PC110 or a Freescale 1ghz iMX51, and you "reuse" an EXISTING case design rather than have that massive $100k case design NRE. it's quite a lot, but the most expensive component by far and above is the LCD [estimated $80 in 1k volumes: it's approx $55 by the time you get into mass-volume 100k pricing] and the next the battery [estimated $30 for a 50Wh battery, i.e. providing approx 7 watts for 8 hours continuous].
approximate individual component breakdown, for 1k volumes, it works roughly like this: CPU $25; DDR2 1gb $18; Case+Keyboard+Mouse: $15; 11in 1280x768 LCD $80; 2gb NAND $6; PCI-e WIFI $9; LVDS, USB-Hubs and other bits: $15; PCBs: $5; Battery $30; SATA in the form of a 375mhz TI AM1808 ARM CPU (not kidding! i haven't found a cheaper way yet!) $9. the irony of adding the AM1808 is that you'd get an extra 10/100 ethernet, an extra high-speed USB-OTG port and two extra SD/MMC interfaces thrown in for free.
there _are_ other ways to get cheaper than that, i am sure. such as using a $32 8.9in or 10.1in 1024x600 intolerable screen; using a USB2-to-SATA IC such as the GL830 from genesys logic which entirely defeats the object of having a 3gbit/sec hard drive interface if it's bottlenecked through a 0.48gbit/sec one; and by using an $8 14Wh battery instead of an estimated $30 50Wh one. all of these things are possible, but you end up with a cheap, cheerful, nasty and intolerable device which you'd resent having bought, where in fact the whole point of the exercise is to get something that you're so happy with that you'd pay _up front_ (to someone that you trust with your money) in order to get it, and you'd recommend it to your friends and family _and_ to complete strangers in the street, too.
so - yeah, you can shave nearly $80 off that by going down to "lower quality" components - but do you _really_ want to? maybe you do. if so, please say so.
last thing: anticipating potential questions, such as "why are you bothering with this? surely there are existing machines, right? surrely there _must_ be something out there, already, such that it is necessary to avoid this insane exercise??" to which the answer is: if you are happy with $300+ and you are happy with only an 8.9in or a 10.1in 1024x600 LCD, suuure, there's no need - but the whole reason why i'm advocating that people band together in order to get group purchasing power is because i don't believe that you find 1024x600 LCDs, or those kinds of screen sizes, to be acceptable. we're free software people: we're _instantly_ outside the mass-market "box".
but - to put your mind at rest, here's a list of what i've been able to find so far. if anyone finds anything different or better PLEASE TELL ME, and if it's better rather than different i will a) shut up b) go buy it myself :)
* CT-PC89E http://elinux.org/CT-PC89E. new machine: sadly, the software suppliers (to chitech) are GPL violators, thus making chitech GPL violators as well (whoops). machine is $102 in 20k volumes, FOB (i.e. direct from factory). 667mhz S3C6410, 8.9in 1024x600 LCD. CPU is upgradeable (on an SO-DIMM)
* Norhtec Edubook http://www.norhtec.com/products/gecko/index.html - $200 retail. 1ghz x86 SoC, 512mb RAM, 8.9in 1024x600 LCD. CPU is upgradeable (on a Module board)
* Leemote Loongson hackable-devices.org/products/product/lemote-yeeloong-8089- a-notebook/ - $EUR 300 retail. 900mhz _64-bit_ MIPS (cool!) 1gb RAM, 8.9in 1024x600 LCD. Leemote also do a 10.1in but again, it's 1024x600.
* Lenovo Skylight. Qualcomm 1ghz Snapdragon, 10.1in 1024x600 LCD, includes built-in 3G. rumoured retail price: $500. can you get one, retail? no - of course not.
* The Always Innovating "Touchbook". 600mhz OMAP3530, 512mb RAM and, yep, you guessed it: an 8.9in 1024x600 LCD. pricing $299 as a tablet and an extra $100 for the keyboard plus extra battery. utterly cute machine, but... it's yet another 9in netbook with an unusable screen...
* Fabled Mythical Pegatron Netbook: 1ghz iMX51, 10.1in 1024x600 LCD. price estimated to be $150 to $200 in volume *over a year ago* - and _still_ nobody's buying, because ... nobody's selling.
so, those are the machines i know about. if you like them, great, go buy them (in quantity 1 or quantity 20,000, it's your choice). if not, and you'd actually prefer something with a better resolution screen, big battery and a decent keypad and trackpad, _do_ say so. also please bear in mind that the approx $212 BOM doesn't include shipping, tax, customs duty, build costs, engineering time or the cost of an SATA hard drive, but it does give you a rough idea of what's involved.
i think there's a lot more people out there who'd want something like this - but how many? we just won't know, until people start saying so.