The Free Computer part 2 (We can do it and we should do it)

Short URL:


The Free Computer is not a dream but it is a reality waiting to happen. All the elements are there we just need to gather them together. There has never been a better time to talk about the idea of the Free Computer than now. Last week I wrote a blog entry about the idea of the Free Computer and where we stand today. I think it's time to clarify my position and set out exactly what I mean, and where we go from here.

I was really astonished to see how many people were energised by the idea of the free computer. It seems to me that perhaps it’s time to do something about it. Tony Mobily this week in his blog GNU/Linux and its closing window of opportunity with OEMs speaks about how GNU/Linux seems to be missing out on potential opportunities because OEM’s are reluctant to sell or support GNU/Linux on their products. I agree with Tony when he says it’s because the conditions seem to be stacked against GNU/Linux (and by extension other free software) and to be honest I don’t hold out any hope for an OEM change of heart anytime soon. Unfortunately these companies are very conservative and not very courageous by nature. Their businesses are built on cost reduction not innovation. This is probably one of the reasons that so much computer hardware is so poor.

In order for the GNU/Linux or *BSD community to be more visible to the public it needs an OEM or Computer company to manufacture a product that works flawlessly, one that does more, does it better and does it with less than any other computer on the market. Such an OEM or company does not yet exist.

Perhaps we need to create one.

The OpenOEM (if you will) would be responsible for defining the specification for a laptop and/or desktop system. They would source the parts and certify them for use on GNU/Linux distributions or *BSD systems. By virtue of being able to order more parts at a time this OEM would be able to encourage component manufacturers to release specifications for their devices enabling the free-software community to write drivers and in this instance the driver code would pass into the free software world. With Open Hardware; if a person wanted to add some new function or fix something that was broken in the driver it would be a trivial exercise and the whole community would benefit as a result.

The GNU/Linux and *BSD community needs an OEM to manufacture a product that works flawlessly. Such an OEM or company does not yet exist.

However, the biggest need right now is, as Tony says, the $950 (or less) Laptop that just works. Until we build it ourselves I cannot see any other manufacturer supplying one. Right now there is no market for a product that does not exist, it's demand is unregistered and therefore it does not feature in the OEM's plans.

In my previous blog, I cited Apple as an example of how to ensure the quality of the user experience and they do this by controlling the ‘Whole Widget’ This allows them to optimise the OS and applications for the hardware in question. This means they can guarantee performance and behaviour by limiting it to the hardware they have certified to run correctly. This gives them a huge advantage when it comes to a solid and well engineered product. They really do 'just work'.

We have to do the same ourselves. We **must** do the same.

So how are we going to do this?

I would like to propose that a forum be set up with a wiki where discussions can take place as to the best way to achieve these objectives. I have already received feedback from some people who are interested in making something happen along these lines but we need more people to help and many questions will have to be asked and answered.

What type of architecture would be the best choice? What components are open already that that we can use. How will we design things like power control? Should we build on a distribution like Ubuntu as Tony suggests or build our own distro? Will we create groups within the forum to work on hardware integration, software integration, aesthetics and ergonomics? How will we agree on the final specifications and how will we organise manufacturing. If the OpenOEM is to be successful at all I believe that issues such as these would have to be addressed as early as possible.

How should such an entity be organised; as a trust, as a corporation (as in the traditional definition) as a company or as a Co-Op? How should it be administered and where should it be based?

As a community of software and hardware users, I am convinced that the expertise exists here in order to do this thing. Of course, it will not be easy, but then the important things never are. And I truly believe that this is something worth doing.

I think it can be done and I think it should.

Consider this as a call to arms and a statement of intent, but without the community's participation it will be as nothing so please let me hear your opinions and suggestions.

Dale O'Gorman

P.S. There are many noteworthy projects out there right now such as The Open Bios Project or The Open Graphics Project and even The OpenCores Project, which highlight how much can be done independently.We need to be able to provide a framework to bring these kinds of projects together and maybe the OpenOEM could help to do that. There is a great resource on the Free Software Foundation website which lists hardware already deemed to be open at Hardware Devices that Support GNU/Linux



Anonymous visitor's picture
Submitted by Anonymous visitor (not verified) on

Would it be possible to work with a vendor of parts like newegg to put together an "it just works" kit for a Linux PC? I mean, most people build Linux machines from parts, anyway, because they don't want to pay the MS tax. Make an official list of parts known to work with Linux and make a PC out of them. If people don't know how to plug the parts together to make a PC, a local PC repair shop probably could. This way, instead of trying to guess which of the 10,000 mobos, processors, etc work with each other, you could click one of five or so Linux configurations (cheap, gamer, server, programmer, etc) and your cart would be filled with the right parts. Real OEMs would either wise up or be sidelined.

Anonymous visitor's picture
Submitted by Anonymous visitor (not verified) on

It has open hardware specs, it has open firmware specs, it has open scope specs. It has industry backing it. It has small end product builders. What it doesn't have is a laptop and a big scale end product production facility.

Anonymous visitor's picture
Submitted by Anonymous visitor (not verified) on

It's an ambitious project and definitely a worthwhile goal. But, it is going to cost a LOT of money to do this! The reason that the OEM's are able to pump out $450 laptops is because they are doing them on a very large scale. One thing I have noticed about the open source community is that they love anything that is free but, when you ask them to fork out money or bust their ass to do something, suddenly there are only a small handful of volunteers. My prediction is that you'll have 10,000,000 people tell you they are interested in a free computer. You'll have 200,000 who sign up for your mailing list. You'll have 10,000 who participate in your forums and mailing lists. You'll have 2,000 hard core dedicated fans. You'll end up with 500 people willing to put down cold hard cash for a product. And, you'll get 10 guys willing to actually do what it takes to write software drivers and make it happen. I started cracking up when I saw your comment: "I was really astonished to see how many people were energised by the idea of the free computer." If you gave away free cars or TV's they'd be energised about that too. I hate to be so negative about all this but, I've worked with the FOSS community too long and realise that there are a lot of freeloaders about. But, good luck with your project -- I hope it works out for you.

Anonymous visitor's picture
Submitted by Anonymous visitor (not verified) on

Finding capital is going to be key. Try getting in touch with people that have mounds of money and vested interest in desktop linux; Mark Shuttleworth sounds like just the right guy. He also influences Ubuntu; a great distro.

Regarding distros; don't roll your own; that just adds to the complexity and most people that would get a machine from here already have their distro of choice. Make it so existing distros can certify against the hardware and make support deals with the distribution. If someone orders an Ubuntu machine, they are directed to Canonical support; RedHat machines are routed to RedHat support, etc.

Regarding hardware; I've had really great experience with Areca RAID cards. Not laptop hardware, but someone out there will be in the market for a workstation/server; these cards are awesome! Areca has their drivers in the kernel too, even better.

Anyway, just some ideas.


Anonymous visitor's picture
Submitted by Anonymous visitor (not verified) on

from organon at

You are aware that there is something like that already in existence? It has industry backing and all the open specs for hardware and firmware. And there are companies putting out products for it, namely genesi with their pegasus line. They don't have a laptop though.

Anonymous visitor's picture
Submitted by Anonymous visitor (not verified) on

That's not quite it. Certainly not what I was thinking, OpenOEM would actually do reselling in my mind, would partner with other resellers.

Geoffrey Lehr's picture

I'd vote for a co-op, partially because that would be with the spirit of the community, and to reduce the likelihood of a first-to-market hijacking the project. It may also keep the prices lower.

Also, a distribution designed to really showcase the hardware would be nice, but it would probably be best to stick to an established one to start (so as to not be distracted).

The other day, I asked a friend if he'd be willing to buy something like this, even if it cost more. His answer was an enthusiastic 'yes!' I agree.

Anonymous visitor's picture
Submitted by Anonymous visitor (not verified) on

My God yes this has been in the back of my mind for a year or so and must be in many others.

I won't start it but I will buy it!


Anonymous visitor's picture
Submitted by Anonymous visitor (not verified) on

why not just run a certification program?

Anonymous visitor's picture
Submitted by Anonymous visitor (not verified) on

The never ending increment seems to just confuse the consumer.
Most consumers think there computer is slow because they have too many songs on the hard drive.
There should be yearly models, this way each model becomes a platform to tune software to.

And new technologies can be evaluated before being introduced.
The community could evaluate hardware for the year and pick the best products that will fit the needs of the year to come.

Releasing end of august to catch both back to school, Black Friday and the Christmas rush.
It would be easy to say of you have an openOEM 2006.. yeah it's time to upgrade.. or perhaps trade in?

Maybe subdividing into General User, Content Creator, Gamer would be a good idea.

I agree with the co-op. Get allotments of interested buyers.. no unsold systems. Make the whole co-op open.
We pay our employees this much which factors down to being this much per unit on this allotment.. the parts cost this much.. shipping this much.. the works.

Anonymous visitor's picture
Submitted by Anonymous visitor (not verified) on

As someone who has run a small computer company, I am painfully aware of the challenges involved in assembling, selling, and servicing hardware. The market is so utterly saturated that in order to get any notice you have to do something that is 10x better than what other people are doing. The $100 laptop is an example of an idea that was powerful enough to get real support, but of course it only works because of considerable funding from outside sources. I believe that if you told this idea to anyone who has been successful in the computer hardware business they would tell you that it would be a tremendous amount of work with a low probability of success and very limited benefits. If you believe that you can achieve enough volume to be able to influence hardware OEM's then you should just make a few calls and see what volumes would be necessary.

I am not trying to be critical for the sake of being critical; I just don't like to see people waste their time trying to create a revolution in an area in which they have absolutely no understanding.

If you want to make Linux more compatible with existing hardware then you should support projects that certify particular configurations for Linux. You can then either build the configuration yourself or pay someone a modest amount to build it for you. Hardware OEM's will continue to make their drivers more widely available if more people are using Linux in general, but not because some relatively small manufacturer is asking them to.

I believe that you would find the number of people who would be willing to pay more for a Linux box is extremely, extremely small. I have sold Linux machines to people before and it is almost always the case that they are looking for the lowest price possible.

It is good to dream, but it is important to know what you are talking about before you try to turn a dream into reality.

Anonymous visitor's picture
Submitted by Anonymous visitor (not verified) on

linux is not ready for the desktop market.

Anonymous visitor's picture
Submitted by Anonymous visitor (not verified) on

oh shut up. for mom and pop who go around checking email and websurfing it's perfect. My mom does that, listens to her music and uses the "other version of word" on.

Pandu Rao's picture
Submitted by Pandu Rao on

Yes, an independent OEM is one of the few things standing in the way of widespread adoption of Linux on the desktop.

Any solution we come up with is going to require capital to execute. Steven J. Vaughan-Nichols puts it in the range of 50-150 million dollars [1].


The FOSS community has thus far focused on creating good software. With OpenOEM, our challenges will be those that are traditionally faced by businesses e.g: procurement, operations, costing, distribution, etc. i.e unless we come up with a simple, sustainable operation-model (not business model).

I suggest we solve the harder problem first. i.e, figure out how we're going to raise 100 million dollars. Donations/grants from organizations, micro-donations from individuals, voluntary effort, etc.

That's about 400,000 people donating $250 each. Achievable? Unquestionably!

This is the approximate amount we will ultimately require. Let us recognize that upfront and work towards it. Or else this will end up being a theoretical exercise.

Let's go for it.

Pandu Rao

Anonymous visitor's picture
Submitted by Anonymous visitor (not verified) on

Mark Shuttleworth, the founder of the Ubuntu distribution, has a _large_ amount of capital, which he may be convinced to put towards this. After all, it would be a great way for him to get Ubuntu onto new machines with the guarantee that it will work flawlessly.

Come to think of it, I'll email him about it.

Cam Cope


Thpn's picture
Submitted by Thpn on

Maybe the easiest way to achieve this goal is to work within the system at first. A co-op that is a HP channel partner could leverage existing economies of scale to produce a product that is close to achieving our ideal, out of the starting gate.

Without stockholders siphoning off profits, the co-op would have the working capital to eventually become an OEM. Meanwhile, if this venture is truly viable, the importance of the co-op as a business partner would allow us to influence HP and its component suppliers to open whatever remains proprietary.

This approach represents the lowest barrier to entry for the OpenPC that I can imagine. With headquarters in Grenoble (Is HP's PC unit still headquartered there?) and distribution units in Europe and the US, to start, the venture could eventually be expanded worldwide.

Everyone seems to agree that any OEM that ships Linux should just pick a distro and run with it. So, load Ubuntu with the KDE libraries, so the widest possible market is served.

Start with a rock-solid foundation and build on that. Don't try to do it all at once. Expand as you go. Keep customers happy.

Time is wasting.

Anonymous visitor's picture
Submitted by Anonymous visitor (not verified) on

There seems to be some confusion -- are you proposing a free-beer laptop, or just a laptop that is certified to run free software?

The OLPC is already a laptop that uses nothing but free software. See They are in the middle of heavy hardware and software development, so their laptops are not yet available in the market -- and their first priority is to get them into the hands of third world kids. After that's done next summer, there will be a way to buy these laptops in the First World.

While the OLPC is not for everybody (it's slow, no disk, small screen), I think that some of its concepts will migrate into high end laptops. These include: much improved screens, turning the CPU off invisibly many times a second, use of free BIOS to enable innovative hardware/software collaboration, operation without disks or with disks spun down most of the time, mesh wireless.

sitor's picture
Submitted by sitor on


I just ordered a new PC without OS. Did that at a shop that could guarantee me that all parts would be linux compatible (I don't follow the hardware scene much). If a OpenOEM would have existed, i would probably have bought it from there.
Now I think a project like this is possible, but should not aim too high at first. Maybe a real OpenOEM is a bit too high. I think though that there are other ways to accomplish already a lot. What I would propose is:

1. Start by working with an existing local PC vendor that already has some knowledge on PC's with linux. Collect information on HW openness and together with that local PC vendor create a specification for a few models that are as open as possible. Market these models via an "" website. Do that with HW that is on the market and as open as possible (best in class pieces). Publish the configurations and openness of each component (full open HW specs as the ideal, but just HW with 100% feature support through OSS drivers OK for a start) on, so that other local vendors can reuse the specs.

2. List on all vendors that sell PC's with OpenHW specs. Distribute stickers to these vendors with a OpenHW certified logo.

3. Contact existing linux distributors (or their communities) and ask them to make their next version to work out of the box with the components chosen by, and full configurations.

4. List on all distributions that are certified to work well (per configuration).

5. Organize yearly elections of most open HW manufacturors (in different areas)

6. Once something is running and there is a basis, you can try to confront existing OEMs and request them to contribute or at least create PCs build with OpenHW certified components. In case they comply, give them the right as well to use the OpenHW certified logo.

This way you can start with low costs, local initiatives can hook in (e.g. the shop where I will get my PC this afternoon here in Belgium) and it becomes a real community initiative. Even in locations where there is no vendor that has adopted the OpenHW configurations yet, a user can print out the specs for a given configuration and request it to be custom made by any local PC vendor.

This will link existing OSS communities (via the linux distributors) with the local hardware vendors. Local HW vendors will get to know linux. More people will be asking for HW components that rank best in terms of openness, which will put openness in the picture of HW manufacturers. Any existing OEM can build upon the OpenHW specs and receive OpenHW certification. This might be a project with very low funding requirements (cost of a website and the community effort) but with a much higher chance on success compared to trying to start of a new OEM yourself.

Kind Regards,

Sitor (sitor-at-advalvas-dot-be)

Anonymous visitor's picture
Submitted by Anonymous visitor (not verified) on

It seems to nme we are either talking about and Open OS that runs on any platform or an Open Hardware where any OS runs on it. Either way its a tough Job. MS Windows won't operate on Apple hardware, and Apple OS won't work on Windows hardware, the same for Linux.

But with OpenHW, any OS should work on the platform, including MS Windows if necessary.

I just would like to see a list of components that the Hardware contains so I can make a good judgement when I want to install my OS.

Another idea would be to make a spec. of OS components required to make the Hardware work.

I know that these exist in one form or another on the internet, but some organisation to pull these together would be great.

Author information

Dale O’Gorman's picture