Open source... hardware? Someday, say ETech attendees

The LinuxBIOS project is doing quite nicely, but it's just that: a BIOS -- targetted to run on existing hardware. They don;t design new hardware themselves.

> They don;t design new hardware themselves.

I understand... And your point is???

quote:: unlike software, the cost of such designs is fairly prohibitive, especially when it comes to prototyping. ::quote

Hardware prototyping is a fairly expensive process, especially when one is using multi layered boards like computer boards, The most expensive part of the operation is not the components themselves, but tooling up to actually build a prototype that will reflect the usage requirements of the final product. (I believe I know what I'm talking about here, as I used to design production lines (Methods Engineering) and work in QA for an Electronics manufacturer in NZ)

> I understand... And your point is???

My point is that LinuxBIOS isn't and open hardware project in the sense that the Open Graphics project is (http://wiki.duskglow.com/tiki-index.php?page=Open-Graphics).

You're right, Sander, but I believe there are other projects geared toward hardware. The FSF/GNU Project should have a link somewhere to some Free/Open hardware projects.

@sander

Well yeah, but it is something that will (or at least could) have a more immediate impact. Hardware solutions??? I'm somewhat doubtful...

jimf,

You speak to exactly why I started this thread. What we are talking about likely will not become a 'critical' issue for some time but if there really isn't much past a BIOS and Graphics card in the works what will happen?

I am not trying to be combative in any way, I am just proposing the question..

What are we going to do when the day comes that we have no options left for us to run non-propitiatory software?

What tracyanne says is true, it takes a serious amount of resources just to make a prototype let alone start production on something and have the level of quality and consistency needed.

> What are we going to do when the day comes that we have no options left for us to run non-propitiatory software?

Well, China's got a working processor and seems to be making their own machines. Whether they'd be any better I can't say.

> China's got a working processor

Why am I not reassured...

> Why am I not reassured...

It probably has something to do with your limited faith in governments, but that's just a guess. :)

I've recently been pointed to these:

http://www.opensparc.net/

http://www.gaisler.com/cms/index.php?option=com_content&task...

...which are the OpenSPARC systems.

Unfortunately, I can see "demand" being the real problem with the OpenSource CPU/Architecture. Could speed keep up with what AMD and Intel produce?

Would the difference in price, because the information it is made from is non-proprietary, make up for the difference in speed/features?

Personally, I prefer features in software and efficiency in hardware. It is much easier to update software when a new technique increases the efficiency of a process than it is to forge new silicon.

> ...which are the OpenSPARC systems.

Absolutely! Nice processors, I'd choose one in a heartbeat. Now all we need are motherboards, controllers, etc, etc to get it off the ground, and all at a price that's affordable to the average consumer. I'd really like that to happen. I 'really' wish that would happen. But, I seriously doubt it will.

For true open hardware to become the norm of the day, the entire manufacturing process of computer parts needs to be revolutionized. What you need is some kind of machine that can take a design and produce a product. They are already working on printers that can print circuit boards with metallic ink (e.g: http://www.engadget.com/2004/11/02/epson-prints-1mm-thick-ci...). Someone needs to come up with a way to do the same thing for integrated circuits. The first one who comes up with the technology to do so is going to be a gazillionaire, even if the machine can only make IC's in the micrometer scale instead of the nanometer scale (E.g. Pentium 1). Such a machine, combined with projects such as Open Cores (http://www.opencores.org/) could revolutionize open hardware.

Last year there was an excellent series of articles on Free Software Magazine that described the necessity for such machines (Towards a Free Matter economy (7 parts): http://www.freesoftwaremagazine.com/articles/free_matter_eco...). The gist: If we are ever going to colonize space (even as close by as the moon) then such machines are an absolute requirement. A colony cannot rely on shipments from earth to replace broken parts. They must be able to manufacture them independently from designs transmitted from earth.

Quoting:A colony cannot rely on shipments from earth to replace broken parts. They must be able to manufacture them independently from designs transmitted from earth.

I can hear the construction company now.."What, you mean there's not going to be a service contract for this thing?"

Now you know why we will never do much more that fire satellites into orbit.

The only way we will ever get anything going in space is by helping big business figure out a way to make money on it. They are the only ones with enough capital to do it with any amount of seriousness. Which of course means that ABSOLUTELY EVERYTHING will be as propitiatory as possible.

If not, it will continue to be only military personnel looking busy and the ultra rich taking joy rides.

Well, let's hope that the prospect of becoming one of the ultra rich is enough incentive for someone to figure out how to "print" IC's without the expensive master fab process.

Microsoft locks up the PCs with DRM'd hardware. IBM turns their Cell processor server hardware into desktop hardware. Linux blows Microsoft's doors off.

And the problem is???

Come to think of it, I believe I read somewhere that you can add a hard drive to a Sony PS3 and have a killer Linux box.

"The only way we will ever get anything going in space is by helping big business figure out a way to make money on it."

Obviously someone who never read Victor Koman's _Kings of the High Frontier_.

http://en.wikipedia.org/wiki/Kings_of_the_High_Frontier

But really, what would you pay for a OpenSPARC chip, motherboard and OpenBIOS ROM, and an OpenVideo card? $400? Keep to open industry standards for the IO cards and memory slots, I'm pretty sure it could run Linux right out off the starting block.

There are a bunch of custom ASIC manufacturers who could easily make the chips, who cares if they're made in Vietnam? But _finding_ a market for them is the hard part.

Maybe someone like Penguin Computing as an outlet would work.

There's a huge demand, supposedly, for Linux based hardware. Where, oh where, can we find our deep-pocket "Shuttleworth" to jump-start the process?

> Where, oh where, can we find our deep-pocket "Shuttleworth" to jump-start the process?

Oh, I'm sure the Gates Foundation has a few hundred million to shell out... Just tell them it's a start-up for poor starving penguins in Antarctica... No?... Don't think they'd buy that huh? Well, it was worth a try.

> There are a bunch of custom ASIC manufacturers who could easily make the chips.

The startup costs are still prohibitive. That's exactly how OGP got underfunded back in 2005. They designed a chip, had an ASIC master made and discovered that it had a bug that prevented it from working properly. $1.000.000 out the windows and now the project is struggling to get enough funds for a second try.

These costs simply limit how many iterations you can make of a design. For software it's unlimited. Just change a few lines, recompile and benchmark. This is what lowers the treshhold for contribution. It costs little more then time. That's not possible with hardware design currently. Software emulation is simply too slow and FPGA's are limited in size as well as expensive to buy.

@jimf,

Just tell them that "Happy Feet" was a real documentary.