> Based on Journal X’s practices, my photographs would be
> isolated from the paper, uploaded to Wikimedia Commons,
> and available for corporations who normally pay for my
> images to get them as freebies.
As noted in the article, the journal doesn't have the right to change the license on someone else's work. This is an ongoing problem with liberal copyright licenses in general; I often receive emails from people who ask me to release my software under MIT or BSD3 rather than GPL so they can "relicense" it, and sometimes even several back-and-forth emails are insufficient to convince them that copyright licenses aren't mutable by anyone but the owner.
For the author of this piece, I think the solution is relatively straightforward:
1. Point out to the journal that having a photo in one of their articles does not grant automatic permission to put that photo on Wikimedia. If the journal's software is unable to handle these cases separately, then the software should be corrected.
2. Ask Wikimedia kindly to remove the author's photos from their collection, or at least correct the license metadata. I'm sure the Wikimedia editors would be willing to do this, though they might become unhappy with the journal editor who uploaded photos without permission.
3. If someone uses the author's photos for commercial purposes, contact them and let them know that such use require a commercial license. They will likely be uncooperative (c.f. the various newspapers who like to source uncredited photos from Twitter), but some gentle reminders about copyright infringement's RIAA-engorged penalties should bring them around. If nothing else, they will likely become much stricter about validating ownership before using a photo.
that copyright licenses aren't mutable by anyone but the owner
Well, sort of. Certainly if you've released something under CC0 or WTFPL, any derivative work can be "relicensed" however they want. MIT and BSD and CC-BY have some fairly trivial conditions that must be respected and fulfilled, but otherwise a derivative work can still use a different license.
I don't think many people care about the particulars of open source software licenses except in the context of producing derivative works.
At most journals, the authors of the article must certify that they have the rights to the whole manuscript and agree to relicense it. And I imagine that many biomedical journals wouldn't see the benefit of making an exception for this unusual edge case, or in spending the money to write new software to do so.
I often receive emails from people who ask me to release my software under MIT or BSD3 rather than GPL so they can "relicense" it,
That makes no sense. People are asking GPL software to be relicensed under BSD/MIT to make them usable in situations where the GPL is a legal hassle or impossible to use. I doubt that people even consider relicensing BSD as something else (besides people practically downgrading BSD to GPL).
1. User wants to use my GPL'd library in their application. Their application is released as source code under the 3-clause BSD license.
2. User sends me an email asking me to change my library's license to MIT or 3-clause BSD so their open-source app can use it.
3. I reply with the longform equivalent of "u wot m8?"
4. They reply that if I change my license to MIT, they will be able to "relicense" it to 3-clause BSD and use it as a dependency in their application's build script.
5. I explain that they don't need permission to just depend on a library, since they're not distributing any of it. And besides, if they did include some of my code in their tarballs, they still can't change the license because they don't own the copyright.
I suspect this stems from the same sort of pseudo-religious copyright knowledge that leads to "no copyright intended" disclaimers on YouTube. Lots of people grew up in the era of Napster and copy-pasted MySpace javascript, and never really gave much thought to how copyright actually works.
Presumably, User wants to include your library in their open-source BSD release. If that is the case, what they're asking is correct since a work extending / modifying a GPL licensed code needs to be released as GPL as well. That is the main difference between GPL and non-copyleft licenses.
They could distribute their code without your library, as you say, but that would require more hassle from the end-user, especially if there's some configuration / integration process that is required to make both work together.
My libraries are mostly in Haskell, so depending on them is a matter of adding a line to a standardized build file. I have not yet received mail from a user asking me to change a license because they wanted to actually include code in their works -- it's almost always a misunderstanding of copyright.
Presumably what they want to do is mix the BSD code with their own code and release the whole under a restrictive licence. That's permissible under the BSD licence, but doesn't involve relicencing the BSD code, since that's still available under the BSD licence.
The journal in this case is presumably licensed under a CC licence that permits commercial reuse and derivatives. In that case, they do need the same (or less restrictive) licence on each image that they want to include. Fair use won't be sufficient.
While I doubt they would, in a case like that, they could have included a link to the image instead of the image itself. This is less than ideal in a lot of ways, but it would appear to solve most of the issues.
Comments
For the author of this piece, I think the solution is relatively straightforward:
1. Point out to the journal that having a photo in one of their articles does not grant automatic permission to put that photo on Wikimedia. If the journal's software is unable to handle these cases separately, then the software should be corrected.
2. Ask Wikimedia kindly to remove the author's photos from their collection, or at least correct the license metadata. I'm sure the Wikimedia editors would be willing to do this, though they might become unhappy with the journal editor who uploaded photos without permission.
3. If someone uses the author's photos for commercial purposes, contact them and let them know that such use require a commercial license. They will likely be uncooperative (c.f. the various newspapers who like to source uncredited photos from Twitter), but some gentle reminders about copyright infringement's RIAA-engorged penalties should bring them around. If nothing else, they will likely become much stricter about validating ownership before using a photo.
Well, sort of. Certainly if you've released something under CC0 or WTFPL, any derivative work can be "relicensed" however they want. MIT and BSD and CC-BY have some fairly trivial conditions that must be respected and fulfilled, but otherwise a derivative work can still use a different license.
I don't think many people care about the particulars of open source software licenses except in the context of producing derivative works.
At most journals, the authors of the article must certify that they have the rights to the whole manuscript and agree to relicense it. And I imagine that many biomedical journals wouldn't see the benefit of making an exception for this unusual edge case, or in spending the money to write new software to do so.
That makes no sense. People are asking GPL software to be relicensed under BSD/MIT to make them usable in situations where the GPL is a legal hassle or impossible to use. I doubt that people even consider relicensing BSD as something else (besides people practically downgrading BSD to GPL).
Example:
1. User wants to use my GPL'd library in their application. Their application is released as source code under the 3-clause BSD license.
2. User sends me an email asking me to change my library's license to MIT or 3-clause BSD so their open-source app can use it.
3. I reply with the longform equivalent of "u wot m8?"
4. They reply that if I change my license to MIT, they will be able to "relicense" it to 3-clause BSD and use it as a dependency in their application's build script.
5. I explain that they don't need permission to just depend on a library, since they're not distributing any of it. And besides, if they did include some of my code in their tarballs, they still can't change the license because they don't own the copyright.
I suspect this stems from the same sort of pseudo-religious copyright knowledge that leads to "no copyright intended" disclaimers on YouTube. Lots of people grew up in the era of Napster and copy-pasted MySpace javascript, and never really gave much thought to how copyright actually works.
Presumably, User wants to include your library in their open-source BSD release. If that is the case, what they're asking is correct since a work extending / modifying a GPL licensed code needs to be released as GPL as well. That is the main difference between GPL and non-copyleft licenses.
They could distribute their code without your library, as you say, but that would require more hassle from the end-user, especially if there's some configuration / integration process that is required to make both work together.
My libraries are mostly in Haskell, so depending on them is a matter of adding a line to a standardized build file. I have not yet received mail from a user asking me to change a license because they wanted to actually include code in their works -- it's almost always a misunderstanding of copyright.
Presumably what they want to do is mix the BSD code with their own code and release the whole under a restrictive licence. That's permissible under the BSD licence, but doesn't involve relicencing the BSD code, since that's still available under the BSD licence.
The journal in this case is presumably licensed under a CC licence that permits commercial reuse and derivatives. In that case, they do need the same (or less restrictive) licence on each image that they want to include. Fair use won't be sufficient.
While I doubt they would, in a case like that, they could have included a link to the image instead of the image itself. This is less than ideal in a lot of ways, but it would appear to solve most of the issues.