Building an Open Source Laptop

Bunnie Huang: Building an Open Source Laptop

I discovered the Novena project only recently while watching a video of Bunnie Huang's talk at 30C3 on SD cards. I think it is a fascinating project and I am very curious about the crowd funding campaign. I would love to buy an open source laptop but I fear that it may be too pricey for me. Well, let's wait and see.

What do you think about this project?

 

The problem is there is no ATX Form Factor standard for notebooks. With PCs, the ATX standard insures any ATX PSU from any maker will fit in any ATX case from any maker, and support any ATX motherboard from any maker and all the components will physically and electrically mount and interconnect with ATX defined physical and electrical dimensions and configurations.

With notebooks, there is such competition to make the thinnest and lightest notebook that pretty much all parts are proprietary.

For that reason, you cannot simply go to Newegg, Tiger Direct, Amazon or Best Buy and pick out a motherboard, PSU, RAM, graphics, and case and assume they will all fit and work together. If you could, there would be a booming self-build industry for notebooks - like there is with PCs.

As the article noted,

 

All true and a little sad.
If building a laptop (not even open source) was as easy as building a
desktop PC, many people would have done so and of course, I would have
been very interested as well. But that's clearly not the case, as you
pointed out. I never meant that I was trying to build my own laptop. I
just copied the headline of the article.

I was interested in your opinion on the project and whether you'd be
interested in buying an open source laptop?
Fortunately, Bunnie Huang and Sean Cross have done the hard work and it
looks like the Novena is running stable, with open hardware only. I
don't have a clue how expensive the laptop in the upcoming crowdfunding
campaign will be, but I am very intrigued.

First of all, I've become very interested in using free software only
since using Parabola. Of course, this has something to do with the NSA
scandal as well. But not exclusively. I admire the people working on
free software. Another thing is, that I like that in the prototype a
very nice display has been used but it seems there may be other options
available as well.

Using Linux almost exclusively over the past few months, my perception
towards the hardware used in laptops shifted. I noticed that I don't
need all the latest stuff in order to do what I like. I know I am
repeating myself a lot, but my old laptop is working fantastic with
Parabola. It takes <5 seconds to boot up, the same applies for
shutting down. I think an open source laptop could be my new laptop for
at least the coming 5 years. Unfortunately, I don't have a lot of money
to spend on technical gadgets that's why I am still using a 7 year old
laptop and an even older mobile phone.

 

Open source is great - if there is a demand for it as that tends to drive prices down. This works good for software. But there just is not much of demand for open source hardware.

 

Not yet. With awareness comes demand (well, I hope it does). I think,
people need to be made aware of how things work
(see this article) and what you can do about it. Supporting a project
like this would be a start. Of course, I can't do it if it were to be
ridiculously expensive but even then I'd hope that enough people would
support the project.

 

Or rejection - depending on the potential benefits and downfalls.

Here's the problem as I see it. Dell and HP can go to ASUS, Western Digital, and Micron promise to buy 1,000,000 motherboards, hard drives, and RAM modules over the next year. For that promise they can purchase those components at HUGE volume discount prices. This is how the popular notebook makers can offer brand new notebooks (with W7 or W8.1) for under $300.

And since costs is a HUGE factor in consumer buying decisions, I just don't see open source hardware being anything but an enthusiasts (someone who doesn't care much about costs) purchase.

 

As stated by BB. The cost would be prohibitive.

I would say in an apples vs apples situation you'd pay 3x-4x what the big guys pay.

 

I think you are right regarding costs and that it's the factor in
almost everyone's decision.

Nonetheless, I think there might be enough enthusiasts being interested
in such a laptop. I couldn't afford to pay 3 or 4 times the price of an
equally equipped non-open source laptop but I would be willing to pay
significantly more money because of the effort that has been put into
this project and the philosophy behind it. Of course, there are limits.

It may allways be a niche product for a niche market but I wonder if
this might be a product that could be interesting enough for certain
manufacturers to build a product that is affordable for a wider
audience. Or some kind of sponsorship to reduce the price. I don't know.
I probably sound naive but I just hope that this project won't be exclusively for
those for whom money is no object.

 

The idea behind open source is nice - especially with software because virtually anyone can make modifications to it. But that is just not possible with hardware. Factories have to be retooled to accommodate changes. And if the hardware requires driver modifications, who is going to do that?

There is already a LOT of common code out there - if not you could not use the same hard drive, RAM or even CPU with Windows, Linux or Macs.

 

I am no expert but isn't it also about understanding what's in your
computer and how it works? The article says it should be hackable but
you'd need to be an expert to do that. I am not. What interests me, is
the fact that on the one hand an expert should be able to explain to me
every detail of the open source laptop and how it works. On the other
hand I could just read the manual. If I understood the article correctly
there are no parts unknown.

To be honest, it is the philosophy of being open that interests me the
most. Not being able to understand the blueprints may seem ironic.
I'd need to trust these guys. But I'd gladly trust these guys over the
big companies. The past few months have shown that a lot of the
technical devices seem to have strange parts of hardware (for example
routers).

 

No. You don't need open source to know what's in your computer or how it works. Remember, you don't need Windows. You can use Linux because all the I/O protocols are standardized and well documented.

You don't need to be an expert to be a hacker. You just need the tools which are easily available.

 

Even on linux, in the end, you don't know what's really happening when the asm is interpreted by the CPU. Not without documentation or an open source piece of hardware.

Open source is a step towards opening up the aspects of the CPU that are otherwise kept in obfuscated hardcoded "instructions".

 

I don't see that as a bad thing. One thing open source does is stifle competition, options and even progress.

If AMD and Intel did not keep their secrets secret, it is likely we (consumers) would have only one brand and just a few models to choose from.

 

I really couldn't disagree more. Their secrets aren't in their implementation of mv or cmp or something like that. That gains them nothing. Their secrets are their R&D, working on new technologies and architectures, etc. That's what makes intel fast, or amd cheap, or whatever else they benefit from.

Keeping these things closed really doesn't benefit them financially all that much. I think it's driven much more by fear, and that's lessened quite a lot - that's why graphics driver manufacturers have started releasing so much documentation. It doesn't hurt them to do that, it benefits them.

 

Yeah, we definitely disagree. You are talking dozens, if not hundreds of patents and other trade secrets. And while "secret" may not be the best term, having "exclusive" access to that data (either by trade secrets, patents, or copyright protections) is what makes "intel fast, or amd cheap, or whatever else they benefit from".

If there were always a level playing field, neither Intel nor AMD would have any incentive to continue their R&D to make their products standout among the competition.

And I note the GPU makers release so much info so game developers will take advantage of their unique features - not so AMD can steal (I mean copy) from NVIDIA and vice versa.

 

The documentation doesn't help game developers at all. It is purely to contribute to open source projects that attempt to create drivers for their hardware.

AMD and intel hire others to implement features in the closed source drivers.

Going to GPU's again, the open source projects are built entirely off of reverse engineering closed source drivers. So if a competitor really wants to gain some knowledge of how your driver works, they can just reverse it. They have a platform and access to the hardware. There are no trade secrets anymore.

Nouvea drivers don't even use typical reverse engineering because nVidia forbids it, and they still create a driver that mimics the behavior.

Keeping it closed doesn't keep what they're doing a secret, it just keeps the code a secret. Every feature is pretty well documented and out in the open at this point, if nvidia sees AMD using a new AA technique they don't need the source code they can already take that technique for themselves.

The real reason they don't open up is because they're too worried that they've accidentally breached a patent or that they'll attract unwanted attention to their code. That and an irrational fear that people will steal secrets that everyone already has access to.

 

Sure it does. Note I am not talking about program user guides.

No. The code is not secret. It is copyrighted.

 

Not really. The game developers build their games using exposed APIs with defined behavior. Documentation about the backend isn't going to help much - you can pass a function parameters, and that's it, no matter what. The documentation does, however, help developers of open drivers.

The code is secret. It's only released in binary format.

 

In most cases, you are right but if what you said were universally true, all games would play exactly the same with both AMD and NVIDIA based graphics cards. But some don't. There are games that are optimized for AMD for best play, and there are games optimized for best play on NVIDIA.

 

Yes, I know. I didn't express myself very well. I rather meant the
„obfuscated hardcoded "instructions"“ Hungry Man was talking about than
how a computer works in general.

Thanks for your comments!
You clearly have more insight than myself. The question I am still
thinking about is what would be the benefit for the average (linux) user
when compared to a proprietary laptop?
You could change parts of the laptop more easily (e.g. the display), I
suppose you would need significantly less power, it is more secure (no
backdoors). Being hackable I suppose performance might increase with
time.

@ Hungry Man
What do you think about this laptop regarding security and maybe other
things I didn't mention?
I am thinking about getting a new laptop that should run linux. I am
still satisfied with my old laptop but I would like a laptop with a
better display (+ I could give my laptop to a family member). I read
about coreboot and chromebooks and so called linux laptops (like
system76). And then I stumbled across the Novena project.

There are already speculations about the cost of this thing. Someone
guesstimated at least 1500$. I don't know if I could afford to pay that much.

 

That's a tricky question because many would argue that no Linux user is an average user. So the "average (Linux) user" is typically already more experienced than many Windows users.

 

@Bill,

Scalar vs Super Scalar, etc. There are fundamental differences in the hardware and software for both systems. The games still have their own libraries, their own exposed code. For example you can use PhysX and it'll run faster on nVidia - that doesn't mean any developer has access to Physx source.

@Bal

The system is easier to audit for undefined behaviors, and therefor backdoors. That's the largest implication for security.

Otherwise, I don't know the details of the architecture so I can't say what using this laptop would be like.

Personally I'm happy with my Chromebook's security.

 

Thanks, that's what I thought as well and what interests me the most. I
will have to wait a few months and see what the Laptop will look like
and what the costs will be.

@ Bill
That may be true. With average I meant someone who is not a software
engineer or hacker but knows his way around Linux.

 

Crowdfunding has started:
http://www.crowdsupply.com/kosagi/novena-open-laptop

http://makezine.com/2014/04/02/the-worlds-first-open-source-laptop-makes-its-debut/

 

It won't work. Why? Because a key attractive feature of open source is the ability for users to customize their builds with their own choices of case, motherboard, monitor etc. But the fact is, there are no notebook case makers, for example, for self-builders to choose from like Antec or Lian Li.

That is, because their is no ATX Form Factor standard for notebooks, customization options are, and will remain extremely limited, if possible at all.

For those unaware, the ATX Form Factor standard for PCs ensures any ATX motherboard, drive, PSU or expansion card from any maker will fit and be supported by any ATX case. It also ensures all connectors, voltages, and mounting options will be compatible. Not so for notebooks which tend to be very proprietary.

As we can see by those images, those notebooks are huge (thick and somewhat heavy). And really? $2000 for a slow 1.2GHz notebook that supports only 4Gb of RAM and has a tiny 13.3" screen? It does not support SATA III or USB 3.0 so it is several years behind in the latest hardware technologies. It doesn't even come with a keyboard,

This is all way too little and way (like 15 years!) too late.