"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: 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!
发布评论
评论（1）
事实上，这几乎是一个要求。如果任何人都可以随意更改它，那么它就不再是一个规范：这将违背“提高大量数据提供者和搜索算法之间的互操作性”的整个目的。
达尔克：是吗？我已经习惯了实现定义和临时格式定义以及那些左右破坏规范的人，我认为这不会添加任何东西，如果我决定不继续维护代码，保护将阻碍未来的扩展。我认为商标法可以更好地处理一致性问题，例如违反菲利普斯红皮书的基于 DRM 的 CD 不能使用“CD”徽标。
实际上，你已经习惯了，只是没有意识到：这就是为什么你可以只写上面的三个字母“GPL”，并盲目地假设每个人都知道你的意思，因为 GPL 本身包含完全相同的限制。 （“每个人都可以复制和分发此许可证文档的逐字副本，但不得更改它”）GPL 本身并不是在免费许可证下分发的，正是因为如果任何人被允许修改它，它将失去其意义。
Dalke：你是对的，尽管当我考虑可能性时，我立即想到了 GFDL 的“不变部分”。我要指出的是，人们在许可证授予中做了一些事情，修改了 GPL 的条款，使其成为非自由的，并且我亲自修改了三条款 BSD 许可证，以删除 Berkeley 并加入我的名字，但这些都是狡辩。
是的。从你的问题可以清楚地看出，你非常关心尽可能降低实施者的障碍。那么，如果我无论如何都必须支付专利许可费用，那么免费、开放、免版税的规范有什么用呢？ 必须解决这个问题，最好由在此类问题（包括但不限于开源项目在专利许可方面面临的具体挑战）拥有丰富专业知识的知识产权/专利律师来解决。
其中存在一些非常微妙的陷阱。例如，一个常见的主题是要求根据通常所谓的 FRAND 提供专利许可（或RAND）术语，其中代表公平、合理和非歧视。这听起来不错，对吧？但这里有一个微妙的问题：每份收取 1 美分当然是合理，如果你向每个人收取相同的费用，那么也是公平且非歧视< /em>.除了开源项目（甚至是可自由分发的专有项目）无法强制执行这些条款，因此它们无法实现该规范。
达尔克：非常正确。但对于许可证来说，这是一个很好描述的话题。有大量关于此事的文字、建议、播客，甚至自动许可证选择器。对于规格来说，没有那么多。我确实知道兰德问题，而且我听说过有关其他规范的故事，其中一位贡献者最后说“哦！看看那个！我们已经获得了它的专利。我们很幸运！”问题是我应该为此担心多少。
因此，适当的专利承诺或契约或无论你如何称呼它们，非常非常重要。 （顺便说一下，商标也是如此。）
例如，W3C 最初希望对其规范采用 RAND 许可，但在受到 Mozilla 和 Apache 等项目的强烈抗议后，他们决定采用免版税模式。因此，即使是一个非常关心自由和开放性的组织也几乎犯了一个错误，有可能杀死每一个开源 Web 浏览器、提要阅读器和 XML 解析器。
“即兴”重要的法律决定是人们最终破产甚至入狱的原因。或者至少非常不高兴。虽然在这种情况下前两种情况不太可能发生，但我认为如果您在两年内发现您的规范由于其专利/版权/IP 法律术语中的故障而完全无用，您将会感到不高兴。
达尔克：我知道这个词会很吸引人。 ;)
有些律师事务所专门为开源项目的非营利开发者提供无偿服务；也许其中之一会对您有所帮助。最知名的可能是美国的软件自由法律中心(SLFC)和德国自由法律研究所 (ifrOSS)。
而且你知道吗，ifrOSS 主页上的第四条新闻是关于开放网络基金会协议，这是 开放网络基金会 特别是开放的、非专有的社区驱动的规范对于网络技术。
达尔克：谢谢。我在瑞典，所以我想知道这些资源对我是否适用。看看 OWF，我发现它是基于美国的，但它努力走向国际化，我看到了一件事我不喜欢：归因的要求。看起来他们确实是可以交谈的人。谢谢指点！
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.
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.
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.
"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!