Very interesting article and I'm glad to see an increased interest in this are.
I've been a core member of the OpenRISC project for four years now, and I can tell you that we already have a fully working IP ecosystem of CPU, perihperal cores and software. There also already exists "Redhats of Open Source Hardware", companies that build complete SoCs based on Open Source components. My own pet project for the last three years, FuseSoC, is also aiming to become the RPM (or dpkg/apt-get) for FPGA designs.
For now, I would say that FPGA implementations are the main target though. The crowdfunded ASIC project at OpenCores was a very badly executed and has unfortunately done more to discredit Open Source Digital Design than helping it. That's not to say that ASIC implementations are out of reach. The OpenRISC CPU has been involved in many commercial products such as Samsung Digital TVs, Zigbee ASICs, and in some of the Allwinner SoC. It's also heavily used in academia and was even sent to space on the NASA TechEdSat satellite a few years ago.
For those interested in knowing more, I'm writing on the topic on my blog at http://olofkindgren.blogspot.com/
@AlPothoof: It was my first color TV with a remote control.
When I was a kid, my grandfather had a stroke, so my parents turned our dining room into a bedroom and he came to live with us.
We had a black-and-white TV in the family room and my parents got a second B&W tv for Grandad's room.
There were only 3 TV channels at that time. I was the remote in the family room (my mom or dad would say "Clive, turn the TV to the BBC" (or the ITV or whatever).
My dad made a remote for my granddad -- it was the long wooden handle from a broom -- my granddad could use it with his "good hand" (he wa sparalized down one side) to change chanels by himself rather than asking one of us to do it.
I can remember my granddad's delight when he could do this by himself :-)
@Max: I remember when you coudl buy a Heathkit and build something yourself -- up to and including a color television -- for much less than you could buy one.
I actually built one, back in the early 80's; took me 24 hours. It was my first color TV with a remote control.
I did it for 2 reasons: 1) you didn't have to be a EE to build it (a good thing as I have a surprisingly small amount of hardware background for an embedded software engineer) and 2) you could troubleshoot it later and parts were actually replaceable.
I don't remember why I got rid of it, probably reached a point where I couldn't get parts any more. I do remember having the case off of it in the late 90's (I don't remember if I was troubleshooting something or just cleaning out out the accumulated dust). That would have been at least 4 household moves after I built it so they were quite rugged.
I had not seen that. Thanks! That took serious courage to outlay the capital needed.
There is tremendous untapped potential for optimized low-power signal processing architectures, but if it isn't for a server farm or a cell phone, I fear that an established chip company is probably NOT working on it.
@Jack.L: Even with free labour, when has anyone been able to build anything of late for cheaper than they could buy it where electronics is concerned?
It's scary in a way.
I remember the days from my youth in England when there were a radio/TV repair shops all over the place -- also you could get a TV repair man to come to your house (the doctors woudl come to your house also).
I remember when you coudl buy a Heathkit and build something yourself -- up to and including a color television -- for much less than you could buy one.
Now I look at the stuff aroudn us and think "how can they possibly design, fabricate, box, and ship this and sell it in stores at such a low price?"
@pinaz: ...and wistfully wish that future might enable the same for multi-project wafers using older geometries...
It's not quite in the "affordable by anyone" category, but did you see my column about the guy who designed his own SoC from the ground up in his basement and had it built as a shuttle project with the ASIC company? Click Here to see that tale.
As an engineer who does both firmware and hardware, I've looked at this phenomenon with bemusement.
Programming the blink rate of an LED has empowered the Arduino/Raspberry newcomer, and now giddy with their perceived success, they want to do the same thing in hardware (but misbelieving the next hurdle will be similarly easy).
Tweaking open-source software is one thing, but do we really need a proliferation of alternate, customized microcontrollers each with an instruction set with "remixed" opcode bits!?
Having said that... I've watched with delight as the cost of firmware and PCB development costs have come firmly into the reach of the average user, and wistfully wish that future might enable the same for multi-project wafers using older geometries. I almost think open source obsession hinders this progression; if everyone insists on giving everything away (and providing free 24/7 support to boot), there is no incentive for newcomers to innovate.