Free Software Foundation Endorses Its First Laptop

[chithanh]

I think a (mostly functional) blob-free Computer is almost impossible to achieve using recent AMD or Intel hardware.

Intel systems for instance require one or more signed blobs be loaded into the chipset during boot:

http://www.coreboot.org/Blob_Matrix

http://www.coreboot.org/Binary_situation

Then the laptop will supposedly have 802.11n Wi-Fi, which means either Qualcomm/Atheros AR9271 or proprietary firmware. Good luck negotiating pricing with a manufacturer who knows that you have no alternative.

I hope that this project succeeds though, it would be very nice to get a more modern free laptop than the Lemote Yeelong 8089/8101 and the Lenovo x60.

[stiiixy]

I wonder, is there a small-form laptop that is tougher than the X-series? They're not built like the IBM days, but when I was still sing them in 2009 they were pretty damn good, powerful units and mostly supported Ubuntu 8/9 if I recall, out of the box. I dont recall there being many hitches to them. Like I said, it took over 90 kilos to break mine. Numerous drops didn't do it! Well, mostly =P

[killertux]

Linux as kernel. HTML5 as the front-end. No?

Free laptop, but what is the use if it can't run HTML5 natively?

Goal for free software is not HTML5 it's just freedom. At least HTML5 does not work with youtube because Gnu LibreJS filters non-free JavaScript.

The laptop store itself is also incompatible with Gnu LibreJS even if all JavaScript there is claimed to be free software. Payment option for PayPal might not be possible because it uses non-free JavaScript and with credit cards I am not sure.

Payment options also lacks free software currency like bitcoin.

[killertux]

I

guess they'd appreciate it.


No. All Intel-based Chromebooks ship with a coreboot port that requires a
binary-only component. You could go through the motions to reverse
engineer that, but then you're still left with the Management Engine
firmware that is mandantory on Intel chipsets starting 5-Series. As
you'll note, all x86-compatible Chromebooks are Sandy Bridge or newer.
That firmware is (supposedly) signed by Intel, making reverse
engineering of that part much, much harder.

As for their ARM offers, all of them start with a binary only (and also
signed) initial bootloader - not much better than their Intel designs.


Once thing they "missed" is the embedded controller firmware. Since
there's no documented/reverse-engineered way to change it (even though
it looks like it's updatable), I guess it's fine for now, even for the
FSF. I half expect some FSF member to work on that next.

I also expect hard drive firmware to be non-free. So it's basically same
as Lemote Yeeloong combined with FSF accepted GNU/Linux installed with it.
However EC in Yeeloong can be patched to free software friendly but I
have not tested whatever it works or breaks the laptop.

[powdigsig]

Linux will live on of course

Linux as kernel. HTML5 as the front-end. No?

Free laptop, but what is the use if it can't run HTML5 natively?

[pgeorgi]

I think that I could reproduce what this provides at a much lower price.

I guess they'd appreciate it.

Is there a Chromebook with Coreboot and wireless that doesn't require a proprietary firmware (with an Atheros card, for example)?

No. All Intel-based Chromebooks ship with a coreboot port that requires a binary-only component. You could go through the motions to reverse engineer that, but then you're still left with the Management Engine firmware that is mandantory on Intel chipsets starting 5-Series. As you'll note, all x86-compatible Chromebooks are Sandy Bridge or newer. That firmware is (supposedly) signed by Intel, making reverse engineering of that part much, much harder.

As for their ARM offers, all of them start with a binary only (and also signed) initial bootloader - not much better than their Intel designs.

Another question I would ask the Free Software Foundation is whether there is any proprietary firmware on the system in ROM chips. To me, there is not that much difference between proprietary firmware loaded from the hard disk and proprietary firmware loaded from a ROM chip.

Once thing they "missed" is the embedded controller firmware. Since there's no documented/reverse-engineered way to change it (even though it looks like it's updatable), I guess it's fine for now, even for the FSF. I half expect some FSF member to work on that next.

[CFWhitman]

Seems a Bit Expensive

I think that I could reproduce what this provides at a much lower price. Of course, as has been mentionded several times, a Chromebook with Coreboot could also be converted to an entirely free Linux distribution. Is there a Chromebook with Coreboot and wireless that doesn't require a proprietary firmware (with an Atheros card, for example)?

Another question I would ask the Free Software Foundation is whether there is any proprietary firmware on the system in ROM chips. To me, there is not that much difference between proprietary firmware loaded from the hard disk and proprietary firmware loaded from a ROM chip.

As far as the usefulness of the hardware goes. In my experience hardware of that vintage can be extremely useful, especially if you are running Linux on it. I use and give away laptops at home with similar CPUs and less than 2GB of RAM all the time. If they have at least 1GB, they're good machines for typical use (most either have 1.25 or 1.5 GB). The only time I generally wish for more than 2 or 3 GB of RAM is when I'm running virtual machines. I'm fairly sure Xubuntu, for example, would run very happily on this hardware.

[CFWhitman]

So that's the whole aim isn't it? Linux-exclusive hardware.

If that's what Linux users want, then they better have the balls to admit that 'freedom' was never their goal, and exclusivity is. Lest they forget what is the real meaning of freedom as defined by their own idol Stallman:

If Coreboot limits me from running any other operating system on a machine that is loaded with it, it is effectively depriving me of the freedom to use it for any purpose. Thus the machine that it is loaded on is not freedom-respecting hardware and by extension, Linux-exclusive hardware are not freedom-respecting hardware.

Until your posts stop consisting entirely of strawmen, it will be hard to take them seriously.

[curaga]

This has got to be the most batshit crazy post anyone has ever posted on phoronix. Well done!

Oh, you should read his other posts. That's not his first phoronix account. Hi ParadoxUncreated!

[curaga]

I think the RYF certification is too strict, in that it requires second-degree 100% freedom.

For example, if I were to sell a laptop with 100% free software installed, and said GNU/Linux on my website, I would fill the first-degree requirement. But if the preinstalled distro was Ubuntu, since Ubuntu has proprietary software in its repos, I would fail second-degree.

I belive the cert would be much more useful if the first-degree was all that was required. Even if Ubuntu encourages closed sw by offering it, I the seller do not (except by proxy by installing Ubuntu), and the hw is still fully free with no BIOS blobs etc.

Any FSF people here? Did I understand the requirements correctly?


--

China is currently making some very nice Macbook Air clones with ARM A9 dual-core + Mali 400 + 720p display, for ~100$. Wondermedia has released the kernel and u-boot for a very similar model. You could probably get 24+ hours battery life out of those.

I was considering reselling those with proper Linux preinstalled, with the usual hw tweaks to make it nice to use. But it seems I couldn't get the RYF mark for those if I used Ubuntu, even if everything installed on the device and recommended by me was 100% free software.