My X60s is getting quite old, and I'd like to get another laptop (or netbook or tablet) without all the backstabbing BIOS and firmware that seem to be quite common on most consumer devices; before someone (Cudder) brings it up, I have better things to do than audit a four megabyte binary. I know two of you were working on hackboard, and there's a bunch of FOSS devboard projects cropping up all over the place; I'm wondering if there's an actual usable laptop using one of them somewhere on the horizon.
Also I don't mind if it's only 1 or even 0.5 GB of RAM and a somewhat slow CPU, the most resource-intensive application I ever use is probably eight megabytes and constantly swapping the web browser (because not all websites work in elinks, sadly).
>>7 Only if you have a salary of 100,000u/y+ it is not. Even then, would it be better to have a laptop at < 540€? That way, you can buy more than one and make a Beowulf cluster, if you have that kind of budget.
Enjoy your x86 ghetto
I hope you can really scale on out of order executions, especially visualization, Loongson 3A/2GQ.
>>11 Nobody is disputing there are cheaper options, genius. There are reasons to buy a laptop other than it being the absolute cheapest one on the market.
Name:
Anonymous2013-10-04 17:25
I have better things to do than audit a four megabyte binary
The good: It has an ARM cpu and a tiny RAM but it seems to depend on no nonfree blobs.
The bad: As usual, mainline support sucks.
The ugly: The "community support" consists of a bunch of (unsigned) binaries posted on file sharing sites along with arcane instructions.
Name:
Anonymous2013-11-27 9:38
>>15 ArchLinuxARM works pretty well with it, though. I think I only had to download one binary blob.
On the other hand, the company is releasing a specialized version to act as a TOR node, so I would expect them either to embrace full openness soon or to be corrupt to the point where the hardware is backdoored.
Name:
>>152013-11-27 9:47
Also it looks like the Beaglebone Black will be the first decently usable board (not requiring blobs (except maybe for 3d accel, but I don't care about that)) for which support will hit mainline for both uboot (it's already in) and the kernel (apparently in january or so).
>>16 let's start a business for verifiable hardware design, I'm sure we'll have lots of big business and military buyers just dying to have the assurance that their anuses aren't being proctologized by competitors/political frenemies.
Name:
Anonymous2013-11-27 10:11
>>17 If you want it to be profitable (in a decade or so), a nice way would be to make sure that all the hardware you produce is suitable for use under heavy radiation. The performance gap between [X produced in a manner that satisfies good verification] and [X bought from a big-name distributor, backdoors complementary] is absurd to the point where your sole userbase might consist of Cory Doctorow fans, but if you can make all your components space-ready without too much overhead compared to simply verifiable, you could win over a rapidly growing market.
The performance gap between [X produced in a manner that satisfies good verification] and [X bought from a big-name distributor, backdoors complementary] is absurd to the point where your sole userbase might consist of Cory Doctorow fans,
I'm pretty sure the military and some big businesses would also take the tradeoff. They've already taken and implemented all sorts of absurd measures and policies in the name of state/corporate security, so they'd jump at the opportunity of finally taking a radical decision that actually makes sense.
Name:
>>192013-11-27 12:58
Also a lot of developers working in security-sensitive environments might want to keep the thing that runs the web browser and video player and torrent client physically separate from the thing on which they edit code and cryptographically sign commits/releases.
Name:
Anonymous2013-11-27 13:02
Ooh, I'm getting all sorts of ideas now. What if the computer has two motherboards, one with a fast but not verifiable CPU and components, and one with slower but fully verifiable stuff on it; both of them are enclosed in faraday cages and only one of them is electrically connected to the rest of the computer at any time using a physical switch. This approach could even work for laptops given sufficient miniaturization.
Also I'm not sure whether verifiable DRAM can be produced or whether the form it is currently being produced in is verifiable anyway.