Choosing licensing statement for an open standard/specification [closed]_问答_开发者

Closed. This question is off-topic. It is not currently accepting answers.

Want to improve this question? Update the question so it's on-topic for Stack Overflow.

Closed 10 years ago.

Improve this question

I've started to write a file format specification for a domain-specific data type. My goal is to improve interoperability between a large number of data providers and search algorithms. I want the result to be available for use, patent-free and without distribution fe开发者_StackOverflow中文版es.

I'm looking for advice on which license to use, both for the specification and for the contributor agreement, if I need one.

If this were software then I know enough about the GPL, MIT, etc. licenses to make an informed decision. If this were a straight document then I would pick one of the Creative Commons licenses, likely CC by attribution.

Looking around, I don't find any common license statement or much in the way of advice. I'm leaning towards the one used in RFC (for example, the HTTP/1.1 copyright statement) but that says "this document itself may not be modified in any way" (with exceptions), which is something I'm not used to from developing code under the MIT and GPL licenses. But that restrictions seems pretty common in specifications.

Unlike most documents but like code, specifications can be affected by patent. Is it best practices these days to also state that the specification is patent-free and require any contributors to reveal any patent conflicts they may know of and/or freely license those patents for the purposes of implementing the spec?

Should I require some sort of contributor agreement?

Or should I just wing it, choose the RFC copyright statement (or CC-By-Attribution), and not worry about this?

"this document itself may not be modified in any way" (with exceptions) [...] But that restrictions seems pretty common in specifications.

Actually, it is pretty much a requirement. If anybody could change it at will, it wouldn't be much of a specification: that would defeat the whole purpose to "improve interoperability between a large number of data providers and search algorithms".

Dalke: Is it? I'm so used to implementation-defined and ad hoc format definitions and people who break the spec left and right that I didn't think it would add anything, and protection would hinder future extension if I decide to not continue maintaining the code. I thought conformance was better handled by trademark law, like how DRM-based CDs which violate Phillips' Red Book can't use the "CD" logo.

[...] which is something I'm not used to from developing code under the MIT and GPL licenses

Actually, you are used to it, you just don't realize it: the whole reason why you were able to just write the three letters "GPL" above and blindly assume that everybody knows precisely what you mean, is because the GPL itself contains exactly that same restriction. ("Everyone is permitted to copy and distribute verbatim copies of this license document, but changing it is not allowed.") The GPL itself is not distributed under a Free License, precisely because if anybody where allowed to modify it, it would lose its meaning.

Dalke: You're right, although the GFDL's "invariant section" sprang immediately to mind when I was considering the possibilities. I will point out that people do things in the license grant which modify the terms of the GPL to, among other things, make it non-free, and I've personally modified the three-clause BSD license to scratch out Berkeley and put in my name, but those are quibbles.

Is it best practices these days to also state that the specification is patent-free and require any contributors to reveal any patent conflicts they may know of and/or freely license those patents for the purposes of implementing the spec?

Yes. It is clear from your question that you care a great deal about making the barrier for implementors as low as possible. Then, what good is a free, open, royalty-free specification if I have to pay for a patent license anyway? This has to be addressed, preferably by an IP/patent lawyer with extensive expertise in such questions (including, but not limited to, the specific challenges that open source projects face with regards to patent licensing).

There are some quite subtle pitfalls in there. For example, one common theme is to require that patent licenses be made available under what is usually called FRAND (or RAND) terms, which stands for fair, reasonable and non-discriminatory. Which sounds good, right? Except there's a subtle problem there: charging 1 cent for every copy is certainly reasonable and if you charge everybody the same amount, it's also fair and non-discriminatory. Except that open source projects (and even freely distributable proprietary ones) cannot enforce those terms, therefore they cannot implement the specification.

Dalke: Very true. But for licenses that's a well described topic. There are reams of text on the matter, and suggestions, and podcasts, and even automated license choosers. For specifications, not so much. I did know about the RAND issue, and I've heard stories about other spec where a contributor at the end said "Oh! Look at that! We've got a patent on it. Well lucky us!" A question is how much I should worry about it.

So, proper patent promises or covenants or whatever you call them, are very important. (As are trademarks, by the way.)

For example, the W3C originally wanted to adopt a RAND license for its specifications, but after significant protests from projects such as Mozilla and Apache, they decided upon a royalty-free model. So, even an organization which cares deeply about freedom and openness almost made a mistake with the potential of killing every single open source web browser, feedreader and XML parser.

Or should I just wing it, choose the RFC copyright statement (or CC-By-Attribution), and not worry about this?

"Winging" important legal decisions is how people end up bankrupt or even in jail. Or at least extremely unhappy. While the first two are pretty unlikely in this case, I assume that you will be unhappy if you find out in two years that your specification is completely useless because of a glitch in its patent/copyright/IP legalese.

Dalke: I knew that word would be a draw. ;)

There are legal firms that specialize in pro bono work for non-profit developers of open source projects; maybe one of those will help you. The most well-known ones are probably the Software Freedom Law Center (SLFC) in the US and the Institut für Rechtsfragen der Freien und Open Source Software (ifrOSS) in Germany.

And whaddaya know, the fourth news item on the ifrOSS homepage is about the Open Web Foundation Agreement, which is a license template by the Open Web Foundation specifically for open, non-proprietary community-driven specifications for web technologies.

Dalke: Thanks. I'm in Sweden, so I wonder how well those resources will apply to me. Looking at the OWF I see it's US-based but it tries hard to be international, and I see one thing I don't like; the requirement for attribution. It does look like they are the people to talk to. Thanks for the pointer!