Hacker News new | past | comments | ask | show | jobs | submit login
Building an Open Source Laptop (makezine.com)
209 points by jamesbritt on Jan 10, 2014 | hide | past | web | favorite | 53 comments

from his blog [1]:

"Funny story about travel – I was on my way back from 30C3, going through securty in Frankfurt. The security agent looked at my laptop and asked me what I do. I said “computer engineer”. He then asks, “did you make it yourself?” and I sheepishly say yes. His face lights up and he says, “ah, so you can trust what’s in it! Now that’s taking it seriously.” I was amazed, the guy totally got it. Wish we had more folks like him working airplane security."

[1] http://www.bunniestudios.com/blog/?p=3597#comment-1341727

Is Frankfurt an American town? Surely it must be, since the Airport security employee was quite polite, enlightened and seemingly open-minded! Awww... time for those delusion pills again ... ;-)

but can you trust what's in it after it's gone through airport security?

Awesome. I wish he used an open-source circuit design tool, but at the same time, I understand why he didn't.

I honestly think that might be one of the biggest impediments to open hardware; open-source circuit design software is not good right now. I try to force myself to use KiCad over EAGLE, but it's such a relative pain in the ass.

gEDA (gschem and friends) has a fairly steep learning curve, but once you spend a few days getting in the zone it's a very solid schematic capture and PCB layout tool. Unfortunately I haven't had a chance to try KiCad for anything serious, so I can't really make a comparison. I'm told it's now the better option vs. gEDA.


They're all a pain the ass. Even commercial stuff such as Proteus and Multisim. Then there's Mentor's offering which is barf.

(I did a spot writing some workflow software for an engineering team a few years back).

Honestly, I've always felt that EAGLE was great. They have a severely limited "free" (as in beer) version, and I really prefer that to KiCad.

It doesn't help that KiCad is horrendously broken on OS X, so I have to use it in a VM.

The good news are that CERN has adopted KiCAD for it's open hardware platform and are actively developing it, adding some cool features which may make it more on par with commercial tools: http://www.ohwr.org/projects/cern-kicad/wiki/WorkPackages

Bunnie Huang[1] is like the Energizer Bunny that keeps going on and on. Glad to hear that he kept plugging away at the project.

[1] http://www.bunniestudios.com/

Off-topic, but something I learned...

I thought surely you meant the "Duracell Bunny", however the Energizer Bunny does indeed exist, and apparently although it's the later of the two, Energizer successfully registered the battery-powered bunny trademark in the US and Canada[1].

Here in the UK I'd say the Duracell bunny is the first to most peoples' minds.

Yes, yes, it's Friday!

[1]: https://en.wikipedia.org/wiki/Duracell_Bunny

Haha, today I learned the opposite of what you learned.

"In Europe & Australia the term "Duracell Bunny" has entered the vernacular as a term for anything that continues indefatigably while in North America the term "Energizer Bunny" has a similar connotation."

Believe it or not I actually hesitated when writing that thinking it was supposed to be Duracell in my mind, but when I said it out loud I was more comfortable with Energizer.

The power of advertising and repeated exposure. It works :(

He's one of my hacking idols, along with Fabrice Bellard!

Yes, it's Bunnie's laptop. And:

The positive response has encouraged us to plan a crowd funding campaign around a substantially simplified (think “all in one PC” with a battery) case design...

It looks like all the positive feedback may result in a production run!

Oh my I love this, open source hardware is the future.

Come to think of how Red Hat is doing great even though they are an open source company. Why can't this work for hardware as well?

It works. Arduino, Makerbot, DIYDrones and a couple of others had 1M+ yearly revenues in 2010, they probably have all grown since then. Open source hardware is probably easier to commercialize than software, because it still has to be manufactured by someone.

> Why can't this work for hardware as well?

I'd be concerned about drafting up blueprints and having them copied by Chinese manufacturers.

Though what I'd do is make my first system just standard libre rights on everything, and that should build a platform for me to kickstart / donation fund future design efforts, and when those are done just release those as free as the last without a care if someone dupes the hardware because we already funded the design.

I'm more concerned that it seems like my only choice is to buy from Chinese manufacturers.

Why? Are American manufacturers any more trustworthy? Most electronics products heavily rely on the Chinese supply chain anyway, regardless of where their final assembly takes place.

My feelings are more about autonomy and choice for its own sake than trust. In this context though, I'd just want to be sure I'm supporting the efforts of people who are responsible for developing products I actually want.

I think you're on the right track, because your hardware is going to be ripped off no matter what. It happens to big companies that obfuscate their designs and it happens to open hardware projects too. At least by freely releasing your designs, a legitimate user can manufacture their own parts instead of just the black-market manufacturers.

So The Freescale IM.X platform is well documented, that's very nice, and I know there's no other company offering something even remotely comparable with more documentation.

It leaves an important question though, there's been speak of NSA/China adding hardware level backdoors to stuff. Is the fact that the firmware is opensource enough to be reasonably sure there are no more backdoors?

For example, Intel has put backdoors in the chips that control their ethernet interfaces. Would something like that be defined in firmware, or at an even lower level? Would we be able to find out if it was lower level? Would we know if Freescale did the same?

What you say is true. Bunnie acknowledges that he can't guarantee a 100% secure platform,

"Our Novena Project is of course still vulnerable to techniques such as silicon poisoning, but at least it pushes openness and disclosure down a layer, which is tangible progress in the right direction. While these heady principles are great for motivating the journey, actual execution needs a set of focused requirements."

Somewhat related talk from 30c3 (speaker is a bit dull but the topic matters): http://media.ccc.de/browse/congress/2013/30C3_-_5529_-_en_-_...

Edit: I see the author was also there, nice (the talk I linked is not the one he gave though)

have a look at thinkpengiun.com

From the about:

Our products are freedom-compatible. Meaning they will work with just about any free software operating system. This is made possible by selling products with free software compatible chipsets.

Free software is a set of principles that ensure end-users retain full control over their computer. Free software can be used, studied, and modified without restriction.

The chipsets we use encourage community development and user participation. Users can not be locked into a vendor or product, be forced into an expensive upgrade, or have other digital restrictions placed on them.

https://thinkpenguin.com is the address (u before i)

Excellent. What are the keyboards like? Bendy chiclets or something decent?

I appreciate the goal but I'd rather we started with an IP-free core manufactured by multiple fabs and work up from there. The Chinese got somewhere with this by ripping off MIPS but I'm not sure that is the right solution.

Then there's the wireless stack and peripheral space which is even more complicated.

Edit: this was my intention when I went to university in the early 90's but I found sitting in front of a Sun workstation drawing squares (gates) very tedious.

Building IP-free core/wireless/... seems to be orthogonal to their goal. I think they would prefer to use such components, but I don't think they are in the right position to build them. It's enough of a problem that cores with NDA-free docs are not common (Bunnie wrote about it in the article).

Awesome stuff.It's time for a standardized case - a big problem with laptops.The folder is cool though.

Anyone spec the cost of making one?

The cost of making one it probably several thousand dollars, not counting the year and a half of Bunnie's time.

From then on it depends on how many you want, and if you can find an assembly house that will take a tiny order for (say) 1000 of them. Even in large runs I can't see it costing less than $1000, partly due to the FPGA on the board. Handcrafted aluminium and leather cases are also likely make up several hundred dollars of the price. That case looks particularly resistant to mass-production.

>"The positive response has encouraged us to plan a crowd funding campaign around a substantially simplified (think “all in one PC” with a battery) case design"

It sounds like the design of the case will be changed if these are going to hit larger scale production, so I don't think that's going to be an issue. It will still likely be expensive though, unless you do a very large run of them.

I tried doing this as well, about a year and a half ago. I gave up relative early in the process, though, as I had hit the upper limit of my electronic design skills, and figured I'd better wait a wee bit more until they would either be on-par with the difficulty of the task, or a good-enough SBC shows up. Based on that experience, I'd venture thinking the price is in the 1000 USD range, with the costs of actually producing the boards (alleviated by more or less preferential access to the required tools) being a significant drag.

That being said, I think it's worth it. I can't think of a single laptop which I'd actually love to use forever, or at least until it's so old that it becomes utterly unusable. The only thing that prevents me from declaring this one as the one I'd like is that OpenBSD is a bit shaky on that architecture IIRC.

If I recall correctly, bunnie's laptop was in the $2.5k range? I could be wrong, but I remember it being very high compared to mass-produced things, which makes perfect sense.

The Raspberry Pi requires a binary blob to function whatsoever, so it doesn't really make sense to use for the purpose of an open source laptop.

I've wanted a FPGA in my laptop for 10+ years. Brilliant.

This project is hopefully laying the foundation for a decent laptop with Cherry MX switches! [1]

Yes, current design is using ThinkPad chiclet but we can still hope, right?

[1] There's also Cherry ML for laptops but I haven't seen those switches myself nor heard about a laptop using it...

It's not a chiclet keyboard. Looks like the keyboard for the thinkpad X220. in my experience it's a very good keyboard. of course I would pay a lot of money for a mechanically-keyed laptop.

woops, yes, I stand corrected, X220 is not chiclet .-)

My T520 is so far the best non-mechanical for me. Possibly the same keyboard as in X220?

pardon my ignorance, but how this is different from non-branded chinese laptops. since those are non-branded i guess they don't have legal rights over it too.

This is a laptop for hardware development and security research, hence the features [0]:

* FPGA on the mainboard

* Lots of digital and analog IO headers connected to CPU and FPGA

* Dual Ethernet


Also, it's not just about licensing, but about having more control over the hardware.

[0] http://www.kosagi.com/w/index.php?title=Novena_Main_Page

The "open source" part is what is different.

"Open source laptop"? Physical objects don't have a source, either open or close. It is rather a laptop capable of running on only "free software" for full functionality.

> Physical objects don't have a source, either open or close.


The schematics for the motherboard and any other boards (like the LCD connector board) can be open or closed. Also, the CAD files for printing the case components on a 3D printer.

Hardware absolutely can be open-sourced.

Wrong. Please see http://en.wikipedia.org/wiki/Open-source_hardware for (much) more information, this is really quite the thing right now.

You might disagree with the definition, if you somehow consider all the information that goes into the manufacturing of a physical object to be something else than the "source" for taht object, but this is how the term is being used by a lot of people, who are actually doing this.

They do have design plans they are built from, so why not call it source?

I am a physical object. I consist of a complex series of 3 dimensional structures, energy flows and chemistry. My source code is a very long one dimensional array of four repeated symbols.

I think it was Ed Wilson who pointed out the 1d DNA leading to 3d structures (perhaps Schroedinger in his book What is Life or Gould in one of the many essays). I can track the reference down if needed.

Meta: I think you are wrong, but I'm surprised at the degree of down-voting.

One of the stated goals was being able to build the firmware from source. As firmware can be considered an integral part of the physical objects that constitute the hardware, I'd say that qualifies it as being an open source laptop.

While accepting the other replies, I do wonder how hardware trust is achieved. The security point about open source is peer review. For a dunce like me, I can pretty much rely on people much smarter than me to review open sourced code. In reasonable time, if there is a iffy routine buried in some bit of software, such people will out it. So, people like me can pretty much trust it, and even if something is missed but later found, again, we can be sure the discoverer will report it, while a commercial organisation might be motivated to hide it.

But, how does that work with hardware? How do we trust all the chips? How do we know there isn't something iffy built in at manufacture? Then, how can those same cleverer than me people discover it? How do they inspect the encased silicone circuit?

So, in that vein, I see the point you might have been making, and actually, I think its a worthwhile question. Because to me, it does look like what you describe, a sort of closed source hardware which runs all open source software. And I personally dont know that that is safe. Or is it?

Open-source hardware and software are analogous in this sense: pre-built hardware is like pre-built software. You can only truly trust it if you built it yourself. However, unlike pre-built software, often you can visually determine if high-order modifications have been made, e.g. additional components, re-routed traces, etc.

You can't really trust silicon, and this leads to a larger point about much of what has been called open-source hardware: many of the open-source boards out there which are really popular are little more than reference designs translated from a datasheet to a schematic capture program for a closed-source blob of silicon. However, this point doesn't escape many people in "the movement," the problem has just been that the technology that meets the users' expectation of performance and capability is out of their financial reach to duplicate. So, the users settle for "re-create" when they can, rather than "trust at all levels". In the majority of cases, absolute trust is not required, because a breech of that trust can result in practically nothing. ("Oh, this AVR chip is back-doored, allowing you to access its 32k of flash, but I haven't provided any means by which a 3rd party could access that via hardware.")

Quote from the original linked article: "Our Novena Project is of course still vulnerable to techniques such as silicon poisoning, but at least it pushes openness and disclosure down a layer, which is tangible progress in the right direction."

Many of the security flaws and known back doors are in the firmware, (see routers for example), which Bunny is resolving by making Open Source. It's probably impossible to eliminate all attack vectors, but by opening the processes up, you can eliminate most of the easier ones.

Applications are open for YC Winter 2020

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact