RANT: A pdf circuit diagram is NOT open hardware

[Rigby]

I'm not sure the part openness even factors in.  You can't open someone else's design, just your own.  You can fully open your design, use proprietary parts, and still be open.  It's the design that's open, not every piece in it.  Some folks will bark and make some noise saying that it isn't truly open, but those efforts are wasted unless they are directed at part manufacturers.  Also, does an open part REALLY matter if you don't have the skills and ability to design, test, and produce an improved part?  Leave that to the part manufacturers unless they get properly stupid about what they're doing.

If you're worried about the part going away or being superseded by some newer version of part that might not work, document why you chose that part, what it does, why it does it, and list things to look out for if the design must be adapted later to use a new part.  This is a hard thing to do because you don't usually have any idea about what's coming down the road, but if you're using an opamp with a particular spec, list that, and list that it's important, so that in the future when your part isn't available anymore you can communicate what is important in the selection of a new part.

Open is nice, but if I hold the design for an opamp in my hands, I will have zero clue at all about how to change it, test it, or get it manufactured in anything close to a reasonable amount of time or money. 

Use an FPGA and/or an FPAA and open the code for those if you're REALLY worried about openness.

[Felicitus]

I'm not sure the part openness even factors in.  You can't open someone else's design, just your own.  You can fully open your design, use proprietary parts, and still be open.  It's the design that's open, not every piece in it.  Some folks will bark and make some noise saying that it isn't truly open, but those efforts are wasted unless they are directed at part manufacturers.  Also, does an open part REALLY matter if you don't have the skills and ability to design, test, and produce an improved part?  Leave that to the part manufacturers unless they get properly stupid about what they're doing.

I'm mostly certain that that comment was about disclosed data sheets. This has, for example, happened on the Raspberry Pi, where people couldn't get their hands on the full datasheet for the main CPU (this isn't a 100%  accurate statement, but you get the idea). I don't know if the Raspberry Pi is an OSHW project (I haven't checked), but if it was, it would be a good example for that statement.

[Rigby]

The Raspberry Pi is not open hardware.  I don't even think the schematic is available.

[granz]

Open is nice, but if I hold the design for an opamp in my hands, I will have zero clue at all about how to change it, test it, or get it manufactured in anything close to a reasonable amount of time or money. 

There is definitely no shortage of opamp varieties in the world to choose from, I was more thinking of fancy motor drivers, RF chips, dc-dc controllers, even some microcontroller peripherals.

Use an FPGA and/or an FPAA and open the code for those if you're REALLY worried about openness.

Well, I'm not personally worried about openness or any such thing, I would be fine using specialized parts and then releasing the design as open source.  OSHW is definitely a buzzword these days, but it seems that a number of the projects are built using some proprietary ICs, some firmware, and not much else.  Often much of the heavy lifting is done by "closed" parts.  My point was more that a bit of perspective on how actually "open" it is can be valuable.

Using an FPGA or FPAA would definitely help if a goal was to really make the design as open as possible.  Actually, the fact that opencores.org exists means that this is becoming more of an option.

[Rigby]

Ah, I misunderstood a bit, as I'm wont to do.  You're right.

[mahjongg]

The Raspberry Pi is not open hardware.  I don't even think the schematic is available.

No it raspberry PI isn't open hardware, the raspberry PI foundation never intended it to be. Nevertheless, schematics (.PDF) of all three current versions can be found here:
http://www.raspberrypi.org/documentation/hardware/raspberrypi/schematics.md\

[miguelvp]

You just said "pdf file" run for the hills! 

[Rigby]

You just said "pdf file" run for the hills! 

I don't get the PDF hate.  I like PDFs.  Beats the hell out of Word or Open Office, if you're interested in read-only stuff.

[miguelvp]

Wasn't about you not even about the OP, nobody is happy no matter what the format.

For example, Intel Galileo has pdf schematics and an allegro brd file for the 6 layer layout. Is it open source? well, you need to purchase allegro to generate the Gerber files, or you will have to painstakingly redo the schematic from the pdf then use your layout software to mimic what the Allegro Free Physical Viewer shows if you want to reproduce the board.

https://communities.intel.com/community/makers/documentation/galileodocuments

So even if all the data is there, people will complain that is not Open Source Hardware.

But some PCB manufacturers will take in the Allegro files but you can't modify it then, at least not easily.

But it's all disclosed so it's "Open" as in anyone can look at the layout and the schematic. But not Convenient which is what people complain about.

Just for reference, we used to get all the schematics of test equipment in printed form so we could repair it. Then manufacturers stop doing that and you had to go hunt or purchase service manuals. But still it was in print not like you could go and do your own board unless you entered the schematics and place the board the same way yourself.

But even if everything is disclosed the Allegro board doesn't have the components on it, so you will have to figure it out from the schematic and the board layout.

Is it enough? I'll say yes, other will say no.

[Rigby]

Maybe an open spec for circuit and board designs is the answer.  Not another open design tool, not even necessarily any code behind a reference implementation of a .ohw (proposed) to PDF converter or something.  For 3D models there are many "standard" formats that are supported by everyone, same for print publishing, audio work, graphics, software, and even things like plumbing and electrical work.  Does EE have any such formal/de facto standard?

I hate myself for saying that.  Open stuff is very often terrible stuff.

[EEVblog]

https://communities.intel.com/community/makers/documentation/galileodocuments
So even if all the data is there, people will complain that is not Open Source Hardware.

They can't complain. That fully meets the requirement for using the OSHW name and logo.

[Rigby]

https://communities.intel.com/community/makers/documentation/galileodocuments
So even if all the data is there, people will complain that is not Open Source Hardware.

They can't complain. That fully meets the requirement for using the OSHW name and logo.

Yeah, that's a lot of doc.  I'd love if an open source project produced this much documentation as cleanly available as this, rather than "use the source, luke" or some other GNU mating call.

[abaxas]

It's all almost incontestable marketing crap.

There you go, I said it. Even the GPL is marketing crap.

[mubes]

I think we need to go back to the philosophy behind 'open' ness as applied to open source and open xyz nowadays.  One of the significant original motivations was to prevent knowledge being locked up and unavailable for enhancement, extension and improvement....somewhat oddly, given the way they're twisted and exploited today, that was also one of the arguments behind the creation of patent law.

So, if we follow the ideas from patent law (unfashionable though that may be) then I would say that a design could qualify as 'open' if 'one skilled in the art could reproduce it without any significant further inventive step'...ok, that's paraphrased, but you get the idea. Not only that, but openness also requires that you don't get your ass sued off if you -do- reproduce the design (and, thus, gives us the argument why the availability of circuit diagrams for HP kit, undoubtedly hugely valuable, could not allow it to be construed as open; only 'well documented').

This definition allows a PDF to qualify as open source provided it gives me enough information to be able to recreate the design, but if, for example, the circuit diagram were of a microwave circuit, with very specific layout constraints, waveguide structures or similar, then it couldn't qualify as open source without Gerbers or at least detailed drawings of the relevant structures.

Open {source | hardware | law | gardening | whatever} should not be about you being able to take someone else's design and make a quick buck out of it for minimal effort, but it -should- be about you being able to stand on the shoulders of giants to be able to see a little bit further.

Regards

Dave

(Edited to remove typos....factual content unchanged)

[SirNick]

it -should- be about you being able to stand on the shoulders of giants to be able to see a little bit further.

There ya go.  That is (was?) the intention of all the open stuff.  Don't (be forced to) reinvent the wheel.  Play on each others' strengths.  A sum greater than the parts.  There's a lot of pointless bickering (in the community) about complying with the philosophy, from both lazy recipients and idealist zealots.  The rest of us just appreciate that we can learn from others, use prebuilt parts to reduce development time, peek in and see how it works, or at least have a snowball's chance of fixing or customizing some thing ourselves.

So, if you've written OSS, yakked in a blog or forum post about what you know, or published a schematic, thanks -- sincerely.  I owe you one.  And hopefully I'll have the opportunity(*) to cast some of my own work out into the community, and it'll help someone else learn or build a thing or two.  That's what open means.

* So what's stopping me?  Well, mostly that I have no intention of publishing yet another 9v battery to blinking LED converter.  If I'm going to release something, it's going to have some value.

[marshallh]

Here is my license

[SirNick]

You should mount the gear with a tamper-proof screw.