Cool, apparently the license agreement for external documents comes in 2 variants: exclusive and non-exclusive – it's simply the only word that changes between these two.

The background is, that, if e.g. another standards body is to sign this agreement, they of course won't transfer exclusive rights.

I've just asked for a guarantee that we'll get the same privilege :)

FR3P281MB21577D698F98B30EB77719A5FA40A@FR3P281MB2157.DEUP281.PROD.OUTLOOK.COM

Just received an assessment from DIN's legal department. Apparently, open source material can be used as a "working basis". I just followed up, to make sure, this also covers for the scenario of "direct inclusion of significant amounts of open source material for which no exclusive copyrights can be transferred anymore" (=our draft). (details in FR6P281MB3789B4989D0C8A1CC22DEEA5E845A@FR6P281MB3789.DEUP281.PROD.OUTLOOK.COM)

few more updates:

1. proposals for de jure standards can officially submitted as many times as desired, but unofficially they're a one-off thing: either the proposal gets excepted by the competent consortium or rejected. It that light, it might be reasonable, to do an updated DIN SPEC first, while getting it reviewed by DIN with the aim of submitting it as proposal for a de jure standard
2. read an updated version of DIN's copyright agreement: it demands all (possible) copyrights "exclusively in advance", which sits in stark contrast to an open collaborative development process (see contribution guide for OHS versions (in the develop branch))

it might also be worth suggesting an official definition to the EC, such as for the term Open Source Hardware

elaborating on the "risk analysis" (my mention above doesn't quite reflect what the term usually means):

• design flaws: make sure that the design (and it's status (e.g proof-of-concept prototype) is described accurately; definitely avoid deliberate misleading descriptions. In other words: your design doesn't need to be perfect, just be honest about it
  • maybe simplified TLR (such as the OTRLs) may help
• instruction flaws: add (general) warnings 1) against misuse of the design and 2) for obvious risks when executing the instructions (risks during replication) and operating the device (risks during intended use)
  • templates may help
• product monitoring obligation: make sure to update your warnings if people report new issues/risks (either associated with the intended use or the combination with other products)

good point, thanks 👍 I'll give this some brain rounds

The above tiers mix IP and part availability. Perhaps pulling this apart with:

• no IP (expired or published), F/RAND, IP protected
• no proprietary parts, at least one proprietary part
• at least one part has 0 suppliers (DIY required), 1 supplier, n>1 suppliers

Problem that I see is added complexity and the status for suppliers may change frequently. But visualizing the dimensions with a radar chart may look cool 😄

The main use case to drive the dimensions: I want to replicate/buy/build-upon an OSH project. I want to quickly make an assessment if the project is suitable for my needs.

What would be other desirable dimensions? :)

moedn (9778122f) at 18 Feb 16:06

moedn (dcfe672a) at 18 Feb 16:01

moedn (4203ea7a) at 18 Feb 16:00

... and 1 more commit

moedn (4f34ab7d) at 18 Feb 15:39

I think having tiers is valuable indeed.

I think this serves the primary use case of the standard: As a potential user of an OSH design, I want to have a quick assessment of potential caveats when using that design.

(Perhaps the tiers could be multi-dimensional, but of course this would increase the complexity and probably is prohibitive.)

see also #90

Suggestion: A tier-based approach regarding the 'openness' of the components.

e.g.:

Tier	Label	Detail	Rationale
1	fully open source	no proprietary components necessary; all necessary components are either standardized (any public standard suffices) or open source (hence compliant with this standard)	unrestricted supply of components
2	requires proprietary parts	at least 1 necessary component is proprietary (e.g. commercial off-the-shelf component (COTS))	component can be reverse-engineered and commercially distributed by independent suppliers; otherwise potentially limited long-term or geographical availability
3	requires proprietary parts under FRAND-licensed IP	at least 1 necessary component is subject to restricted IP, but available under FRAND conditions (as e.g. a standard-essential patent)	distribution by independent suppliers requires a license under FRAND conditions
4	requires proprietary parts under restricted IP	at least 1 necessary component is subject to restricted IP (e.g. a patent)	distribution by independent suppliers requires a (bilateral) agreement with the IP holder

…whereby the assembly itself might also be subject to restricted IP, not just its components

In practice, the standard designation could be something like this: OSH according to DIN SPEC 3105 - Tier 2: requires proprietary parts

Good point 👍 However, I don't find it feasible within the scope of the standard for 2 reasons:

1. For me this falls into the general availability discussion. Even standard parts or semi-finished products (e.g. threaded brass rods) might be unavailable on a global scale (thinking of a team in Chile that wanted to build the OSI² ONE main magnet and had difficulties sourcing the brass parts). We simply can't cover all conceivable manufacturing scenarios and hence never say with certainty that "this OSH can be replicated anywhere". Likewise, I wouldn't make individual rebuilds a benchmark. If we're talking generic requirements for OSH, for some hardware it might just be nearly impossible at this point to enable individual rebuilds (#microchips).
   However, I still believe a label of some sort makes sense, but rather addressing who's in control of the supply than the current supply situation (see my comment below).
2. Availability changes. Components might only be available in bulk today, but individual next year. Or disappear completely (#electronics). A label would require regular target-group- and region-specific assessment.

Perhaps a bit off topic. A problem with proprietary parts is not only that there may be only a single source, but also that this source requires bulk purchases. I believe this is the case for the metal grinder of the blender (I am not talking about the 3D printed part).

Maybe we need a mandatory "warning label" in the standard along the line: "OSH, but at least one part is single supplier with bulk-purchase restriction".

Just had a meeting with Michael Rudschuck from the DIN/DKE OSPO [1], and have exciting news to share :)

1. If we go for an IEC, it's an established standard procedure at CENELEC to endorse it directly. At least in electronics, it's very unusual to start on a national level and go international – it's rather the other way around and so very natural to stay as compliant and up-to-date as possible on a EU & national level. In other words: He believes it's very unlikely that CEN/CENELEC would reject the endorsement of an ISO/IEC for OSH, more so if it targets the CRA.
2. IEC accepts documents licensed under MIT on a regular basis. The only prerequisite is that authors sign a usage agreement with IEC. He didn't know the details, but it seems that there are 2 different contracts – a second one allowing the co-existence of an open source document. I'll get in contact with DIN lawyers to verify this. In either case, because of (1), the PAS procedure at ISO/IEC might be a very viable option, at least as a fallback 👍

--
[1] https://www.dke.de/de/arbeitsfelder/industry/open-source-als-ergaenzung-zur-normung