The brain behind Bug Labs' Lego-like Linux building blocks says we're on the verge of open-source hardware revolution. "[Open Source hardware] will happen. There's nothing stopping it," Bug Labs CEO and founder Peter Semmelhack told The Reg this morning after trumpeting the biz benefits of open-source hardware during a mini- …
why the same?
Why is it always the same "10 years away"?? Why not 8? or 13? Why is it that these people trying to sell vapourware invariably claim that "the big breakthrough" is always "10 years away". Shows a certain lack of imagination that is at odds with the message they're trying to put across really.
sort of the wrong approach
True open source hardware will arrive when those 3D prototyping "printers" drop from $50K each to something the average Joe can afford. I'm talking about the devices that use lasers to solidify plastic one layer at a time.
Already you can find simple & cheap CNC-type machines advertised in Make magazine. That's the sort of thing we need.
This "building block" stuff is great, and I love it, but it's not really something that you can go too very far with.
"And these should work just like software mashups, so you have pieces and parts with standardized interfaces. "
Wow, what a revolutionary idea. Someone should tell Sony
Open Source hardware will not happen, it already happened, quite some time ago http://www.opensparc.net/
Thanks for stating the obvious.
Where to start?
A standardised interface for 240V, 50Hz power connections, let's call it BS1363.
The one thats been overriden by the 230V one last year?
I think I'll patent that one - all in the name of progress!
Sounds like something I read
Think this is what Vernor Vinge has described in "Rainbows End"? Obviously I know that just because a thing is in 'sci-fi' doesn't mean it will happen or even that it's possible, but this one seems a no-brainer and technology is ready for it.
So we come full circle...
Isn't home-brew and DIY kit where Billy G started his empire from? Let's hope we learn the lessons of the past and ensure that something like MS cannot happen again!
"The truth of the matter is that if you do anything moderately successful, you'll be sued."
That's the most succinct expression of what's wrong in the world of IT (and to a lesser degree the world in general) that I've read in a while.
But if you look at someone coming out of school, someone who has an idea for a new gadget and has seen what open-source software can do, they'll be a lot more willing to accept this notion of community-based hardware innovation. They'll say 'Yeah. Shit. I'll give back.'"
"I've this great gadget?"
"What you'll offer me £5million for exclusive rights?"
"F**k the community, I'm of down the boozer."
Roll on twenty years.
Young dreamer now a sad, penniless middle aged alcholic in a bedsit "What the hell was I thinking giving that idea away...why ,oh why, oh why"...glug glug glug, bllleuughh
Lego Building blocks...
you mean those heavly protected, patented,"closed source", trademarked little blocks?
Hardware details are not kept secret to protect IP. They are kept secret to make it more difficult for patent trolls to attack. Sometimes they are kept secret to avoid publicising a bug with a performance trashing work around. Some (many?) of the hardware patents are just as obvious and non-novel as software patents and serve a similar purpose - to keep new entrants to the chip design market in their place.
There is source code for hardware. The most obvious hardware description languages are Verilog and VHDL. Sounds like a good reason for hardware to get the same patent exemption benefits that software is fighting to keep.
Victor already mentioned opensparc. I will add opencores and opengraphics. There are plenty more.
The big problem with open hardware at the moment is economies of scale. It is often cheaper to glue a few massed produced proprietary chips together than to make a small batch of chips using open source cells that exactly match your requirements. If people could buy a small scale 50nm FAB for a few thousands open hardware would take a huge leap forward. If someone has evidence that microfabs are on the way then predictions of open hardware taking off might mean something.
Here's another for your 10 years
Already happening; OpenPandora: http://www.openpandora.org/
As much as I love OS software, Comparing it to OSH is fallacious. The distribution model alone completly screws up the comparison, as does the added risk (Piece of OSS goes tits up, absolute worst is you have to is reinstall your OS. OSH goes tits up, sparks come flying out!). Maybe it will happen, but it's not going to be anything like the open source software model.
Open source hardware is already here, and popular in some circles (look at arduino), and it is now getting much cheaper for indivduals to get high quality low volume pcbs to design their own things. Seeedstudio in china will print pcb's for you at very little cost for very small runs (e.g 3 - 5 boards) as long as you make your designs open source, They print off your boards, as well as some for them to sell in their shop, and will host the designs and documentaion on their website, everyone wins
Some buzzwords are constantly redefined
Some buzzwords are constantly redefined to mean "stuff that we can't do today but we hope to be able to do in the future". One example is the goal of "software reuse". This term is usually used (at least in academia) to refer to not-yet-feasible ways of reusing software. Any software reuse mechanisms that actually work and have been widely used for years or decades are ignored. For example, a procedure call is a form of software reuse (because you can call, that is, reuse, the code of a procedure several times within one program). A class hierarchy provides another form of software reuse. So is a library (of procedures or classes). But whenever you hear somebody talk about software reuse, they dismiss those things are being uninteresting because they actually work and have worked well for ages, and instead they use the term "software reuse" to refer to stuff that doesn't yet work.
Another example that springs to mind is "distributed systems". That springs to mind because I did my PhD in a distrubuted systems research group. Here are examples of distributed systems/applications that work well (and have done so for years): email, ftp, telnet, the world wide web. But if you ask distributed systems researchers what they are doing they will say they are "trying to build distrubuted systems" as if this concept was something new and untested.
I'm not knocking the work of researchers into distributed systems or software reuse; I'm just making that point that no matter how much progress is made in a particular field, people will keep redefining the term so that it doesn't refer to the past successes. This article is similar. It refers to "open-source hardware" and "LEGO-style building blocks" as future goals as if they are thing that have never been done before. In doing so, it neglects to acknowledge the existance of past successes in these areas. One other reader mentioned OpenSparc as an open-source CPU design. Open-source hardware isn't a speciality of mine so other examples don't come to mind. However, there are tons of example of LEGO-style building blocks. Power supplies come to mind. So do RAM chips. So PCI-, SCSI- or USB-based devices.
The real issue is that the person quoted in the article isn't excited about the already existing LEGO-style building blocks because they don't cater for his particular field of interest, and so he wants to see LEGO-style building blocks more relevent to his needs being built. There's nothing wrong with that, but the way it is expressed is extremely sloppy because it gives the completely false impression that are no LEGO-style building blocks in existance today. This sloppyness can be blamed on two people: (1) the person quoted in the article, and (2) the journalist (for not detecting and correcting the sloppyness).
Why do you hate America ?
"If anything, hardware patents are a deeper thicket than software," Semmelhack said. "It will be an issue. But I don't think it will be a terminal issue. The truth of the matter is that if you do anything moderately successful, you'll be sued."
I fail to see how being dragged through the courts could be anything OTHER than terminal for a small startup trying to do its thing the OSH way. He's 'shrugging off' the concerns but that's easy to do when you're not in a court room facing a stream of corporate lawyers.
there's already very successful OSH
The anduino board is a good example of OSH (www.arduino.cc/). There's a small ecosystem around it. It makes adding a CPU to a personal project an awful lot easier.
This'll get interesting...
When 3D printers get chepa, high res and eventually become nanotech assemblers. Roll on the Diamond Age.
Conflating two issues.
There is no reason why modular hardware even needs to be open source.
Look at a PC -- as good an example of modular as we have at the moment.
Proprietary graphics chipsets interface to the motherboard using a standard connector, and a software layer (DirectX, OpenGL etc) provides a further level of standardisation. External gadgets like MP3 players, webcams, scanners and printers connect via standard USB or Firewire connectors, and despite having proprietary hardware inside, communicate using standardised protocols like FAT32, ATAPI, MTP, PCL, PS etc.
Modularisation is all about the interface, not about the internals.
Not only is open source not a prerequisite of modularisation, but I would say it acts *against* modularisation. If you can roll the source into your own product, you don't need to make bulky or time-consuming interfaces -- you can just hardwire it in.
Just look at Linux 10 years ago -- before the rise of the package managers, the majority of people I knew linked in libraries at compile-time. It was only when KDE and Gnome were released that this started to change. Why? Because they were incompatible with the GPL, so you couldn't do a compile-time link.
It was this that forced people to start thinking about run-time linking and dependancy management.
So the same goes for hardware -- if I can roll the existing OSH into the internals of my gadget, why build an interface? And modularity goes out the window. Like it or not, proprietary designs are black boxes that you *have* to build interfaces around.
Open source hardware is here NOW
There are no patent issues so far.
There are no specific open-source hardware licenses. There don't need to be. CC licenses are used AFAIK. Just check some of the many open-source hardware projects on the net (ladyada.com, mentioned this week, or arduino.cc, etc.) See what licenses they're using, if you fancy doing a bit of actual, you know, journalism.
We already use component-based design in electronics. I mean, HELLO?! For instance there's the ATmega168, a very popular component which embeds an entire microcomputer onto one chip. And at a lower level we have resistors, capacitors, etc. OSS hardware has nothing to do with clip-and-play kiddie electronics kits.
What a crappy article.
Open source hardware to come soon, you say?
Soooo... anyone every heard of OpenMoko Neo range? No? Fair enough. There's always projects like RepRap which have full board schematics and firmware code freely available. The arduino system which reprap uses takes the same approach itself.
Then there are things like opencores.org which provide free hardware specifications which you can run on an FPGA.
Of course, most of these have the minor disadvantage that the key components of the system themselves (microcontrollers, FPGAs) are not fully open in the sense of opensparc, and nor do I ever really expect them to be. There's always the risk that the hardware upon which you depend will simply stop being produced, and I do not see home semiconductor foundries on the horizon
Presumably the 'open source hardware' envisaged here is more about standardisation of interconnects and full and open documentation of communication protocols?
Well, good luck with that. I'd be looking at more like 20 years instead of 10.
Is it really open....
Its dumb comparing hardware with software. Hardware development is very difficult to do at home. You need a whole host of speciliased software packages for design and simulation. Then there is an appreciable cost to making prototypes. A substential amount of the Open Source contributions are from individuals people in their own time at home. Commercial companies are generally less likely to contribute with a few notable exceptions (MySQL, Qt, Eclipse, etc.)
So why would a commercial company hardware company provide open source designs with the exception of the component manufacturers.
For example, Intel, Nvidia, Texas Instruments, Motorolla create reference designs which other vendors are allowed to use and modified. In return of course those companies sell more of their precious chips.
This of course doesnt help with interconnect but to some extent, the industry has been getting at this anyhow out of pure necessity.
For instance, we have PCI(x &e) for peripheral board connectivity, we have memory chip standards, we have motherboard form factor standards (ATX), mass storage device connectivty (SATA, SCSI), power supply connectors etc.
Its not perfect but its driven by an industry where no-one builds everything and everyone is reliant on standards. I suggest that the PC is LEGO. Heck, the new Apples conform to the same LEGO standards!
Now what Bug Labs is doing is cool. As they are talking about gadgets. But really all they are saying is buy our hardware , each of the bits you want, and you can write your own software to make a gadget.
I cant see where they released the hardware specs so it can be manufactured by other people. So it isnt "open hardware" at all. Its hardware where you can access all the bits of it in the software.
And even if they did provide the hardware specs to make it trully open, the barrier to entry to create a compatible device is probably great and certainly not hobiest material. So I think they are jumping on the "Open" bandwagon again.
Still I wish them well as it does look fun . Here is the link: http://www.buglabs.net
I see ISA ...
... PCI, AGP, IDE, SATA, USB and much more so what's so revolutionary about OSH ? And why does this come attached with the "Linux" hype anyway other than to impress the fanbois who 'go ape' at the mere mention of "open source" and are positively orgasmic when coupled with "Linux" ?
The same can be seen in the microcontroller world over Arduino hardware and its clones. Nothing more than any other free reference design you can find in a manufacturer's datasheet but tag on the open source appellation and some people's underpants suddenly become a size too small.
Paris : Who knows a thing or two about blowing things up out of all proportion.
Sorry, SoC design started in earnest 10 years ago. The standards for on chip communication are already there. It is a drag and drop approach to new chip design. There is nothing new here.
What he is basically advocating is lots of dev boards that you put together to make a product. Guess what, your competitor will put it all in a board or chip and half the production costs. Who wins?
I'm amused by the way that the example hardware engineer of the future speaks pretty much like Jeremy from Peep Show.
If anything, that's the major problem with a lot of Open Source advocacy at the moment - that it switches interchangeably between the 'good for companies as lowers costs' argument to appeal to The Man, and counter-culture ideologies to appeal to The Kids.
(And indeed, the presumption of every counter-culture has been that it represents a permanent change).
Incidentally, I'd disagree with the assertion that commercial companies are generally less likely to contribute to Open Source projects - I'd wager that the majority of work on significant open source projects (i.e. Linux, Firefox, Apache, WebKit) is being done by commercial developers rather than volunteers.
Of course there are lots of projects out there which don't get any commercial funding at all - probably the majority of projects and majority of open source developers are run by volunteers.
Depends what you mean by "Open source hardware"
Unless you're going to fabricate your own chips, what's the difference between OSH and a marketed chip with a published specification?
"I need a chip to link two bits of kit. Shall I buy the 89 cent chip from Maplin or make my own in my multi-million pound fabrication plant?"
There's lots of OSH out there - I needed a brass bolt 4mm by 20mm and Homebase had loads.
There is plenty open source hardware out there. Look at opencores etc. The hard bit is putting all this stuff together and testing.
Software is relatively easy to make into shareable components because the intereaction between software components is relatively easy to manage.
Hardware is a lot harder to mash up. Add a GPS or WiFi module to a circuit and suddenly you have all sorts of interesting RF interference etc that need other parts of the circuit to be redesigned.
But of course it is far easier to ignore reality and just spout a whole lot of buzzwords at some conference....
Re: Is it really open?
I'm not sure about this open source hardware stuff but no doubt the people who are interested will do better than the mobile telephone makers who spent years trying to get a standard one charger fits all interface, I believe that they have tentatively agreed to implement something within the next few years.
As to the comment.
>Hardware development is very difficult to do at home. You need a whole host of speciliased software packages for design and simulation. Then there is an appreciable cost to making prototypes.
This used to be true. When I started making hardware projects the cheapest way was to make track layouts using etch resist transfers on a copper clad board then throw it in a tray of acid. A strain on the eyes and very laborious. Today, this is almost unheard of as the transfers are all but impossible to find (Mega-UK still do them) even from major electronic components suppliers. However, the advances made are huge, it's easily within reach of anybody to design a PCB using a multitude of free software, I currently use Eagle, then develop it with a UV light box, you still need the acid tray. Also, it's already been mentioned that you can get small quanitity PCBs professionally made for a very low price. Add to this that a single, under one pound, pic microcontroller can do what what before would have taken tens to hundreds of logic chips, counters and discrete components.
Schematics for PIC programmers are available free but the price of a ready built one by microchip and a debugger hardly makes the diy route worth the effort and their free IDE is more than adequate. So no need for "speciliased software packages for design and simulation".
I've not had experience of the ATmega range but no doubt they're just as capable and easy to use as the PIC microntrollers.
As such, hardware devleopment is easily within reach of the individual. The most expensive part of most DIY projects these days is the case. Personally I don't bother with the packaging until the rest of the stuff is working.
- Vid Hubble 'scope scans 200,000-ton chunky crumble conundrum
- Bugger the jetpack, where's my 21st-century Psion?
- Google offers up its own Googlers in cloud channel chumship trawl
- Interview Global Warming IS REAL, argues sceptic mathematician - it just isn't THERMAGEDDON
- Apple to grieving sons: NO, you cannot have access to your dead mum's iPad