The Rust Trademark Policy is still harmful
purplesyringa.moe
purplesyringa.moe
The first basic principle of trademark law in general is that if the use isn't understood by a casual observer [1] to imply endorsement, then trademark protection doesn't apply to that use. Forking the repository to develop a bug fix is the kind of use that's not even going to cross that threshold, so the policy doesn't need to explicitly cover it (although I wouldn't be opposed to firming up the text).
The proposed fix, however, is completely bonkers insane. It basically allows any fork to modify whatever the hell they want and still use the trademark with impunity--i.e., under this policy, something like Devuan could use all of Debian's trademarks because it's "just" a derived work of Debian that strips out all of systemd.
From most of the other points, the author seems to be completely unaware of the legal concept of nominative fair use. It is always possible for someone to use a trademark, without permission of the trademark owner, to actually refer to the good in question. You can always use "Rust" to refer to the programming language supported by the Rust compiler, no matter what the trademark policy says, because that's just how IP law works.
[1] The UK test uses a more poignant phrase: "a moron in a hurry".
They want the right to fork, call their fork Rust and compete with the main repo for users. When asked how having multiple competing repos all claiming to be Rust would be beneficial to the Rust community they had nothing to say.
Being able to call the commands of a fork rustc, cargo etc would allow trivial drop-in replacement in case the official toolchain takes a bad turn (for example with analytics).
Furthermore, being able to use Rust in your fork's name (e.g. g-b-Rust) makes discovering better forks a lot easier, as opposed to the names insanity with Firefox derivatives
There's just little benefit for the public in this Mozilla's fixation with trademarks, open source thrived for decades without it (I know that Rust is not in Mozilla anymore, but the legal culture evidently persisted)
If you're just forking a repo on Github and upstreaming patches, then the likelihood of confusion is minute. Github's users know what a fork means, and if they're the only audience, there's no problem.
If you break the upstreaming link, distribute release packages, market to non-Github-users, and so on, then the likelihood of confusion grows. At some point, a court would probably assess that trademark violation had occurred.
To be clear I believe they have good intentions: they think the book is a good enough specification and they don't want Rust to be sabotaged by bad ideas. But it's an unusual and somewhat untested arrangement: the Rust Foundation's general insistence that Rust = rustc + cargo + authorized backends like GCC seems like a significant technical risk that has gone badly underdiscussed. Too much of the community sneers at the idea of formal specification, but there is no magic spell that guarantees good management in perpetuity at any single organization. Obviously specification has its own problems and risks. But I think the issue has drawn a lot of snitty fights and bad-faith blogs - mostly Rust critics, to be fair - and very little serious discussion.
Trademark law is not that dystopian, it concerns naming products, not everyday speech. Writing (and reading) a trademark use policy with those boundaries in mind seems non-controversial to me.
Creating a fork to contribute a PR is not an issue. If you fork and diverge while still calling it "Rust" then you're probably in hot water. Even then, no one would care unless you amassed a following or notoriety for it, most likely.
Used in what context, though? As far as I am aware, each trademark is only granted for a specific list of market segments. Anything beyond that, and beyond offering products/services, should be (mostly) in the clear.
I think it's a good idea from the Rust Foundation to prevent others from naming threir convention "RustCamp" or something similar, and I don't think they're doing much more.
The most basic rule is that the Rust trademarks cannot be used in ways that appear (to a casual observer) official, affiliated, or endorsed by the Rust Project or Rust Foundation, unless you have written permission from the Rust Foundation.They're saying things like "may not do that" and "cannot be used unless ...", when what they really mean is "we're not granting any special permission to do that", and "cannot be used unless ..., in cases where your local trademark law says you need our permission".
Rust Trademark Policy Updates
I disagree: I think this just specifies that you can't call a different, incompatible programming language Rust.
I think I agree with most of the other points, though some seem like more oversights than intentional restrictiveness. However, I'm not sure they will yield on the event policy. This seems to be a standard that is being set across large open source projects.
Rust doing these things has lead the language to being an example of what not to do in OSS.
Unfortunately they've succeeded in convincing people there is an issue here when there is no issue.
No, that's just what trademark law requires. They were just explaining it.
A fundamental misunderstanding in the first sentence. I stopped reading there.
Please try not to comment on trademark law unless you at least understand the basics.
Maybe the Rust trademark policy page could do with a brief introduction to trademark law to avoid misunderstandings like this.