Change copyright notice of donated copyright code #355
Conversation
OpenGeo/Boundless donated copyright of GWC code to OSGeo in 2014. Update all headers where it can be determined that the code was contributed by OpenGeo or Boundless.
|
(On the whole, this isn't a big deal, but when a script changes 159 files it does warrant a bit of extra nitpicking.) I don't think the ruby script does the right thing. IANAL, but my suggestion would be to replace @author X, The Open Planning Project, Copyright 2008 with Copyright OSGeo 2008 @author X, The Open Planning Project Jody, can you think of anyone at OSGeo that actually knows copyright stuff ? From the readme of GCC, https://github.com/gcc-mirror/gcc Copyright years on GCC source files may be listed using range Random example of a file that demonstrates this: |
|
Sorry about the delay in commenting on this. It's been a bit hectic. When I wrote the script I based it on the way the headers were updated in GeoServer when it was donated. |
|
Also I don't think @jodygarnett will be notified that you asked him a question unless you tag his name. |
| * Copyright OpenPlans 2008-2014 | ||
| * Copyright OSGeo 2014 | ||
| * | ||
| * @author Arne Kepp, The Open Planning Project |
There was a problem hiding this comment.
This look fine, you covered everything that needed to be said.
I usually put the date near the copyright, and spell out the organization name:
Copyright 2008-2014 Open Plans
Copyright 2014 Open Source Geospatial Foundation
Or shorter
(c) 2008-2014 Open Plans
(c) 2014 Open Source Geospatial Foundation
But since you are doing the pull request you can make your own tradition.
|
I should be able to answer any questions you got for OSGeo. |
|
@smithkm thanks for the heads up on the notification @jodygarnett I agree it looks nice, it is a mixed changelog of contribution and ownership which tells a story. But you haven't justified or provided any references to suggest your usage of copyright year is technically correct ? It is hard to find good sources that nail this issue, but here is a concise explanation of copyright year, even if it is from a source of unknown quality: Based on the source above, and the GCC document I linked to earlier, your suggestion would be interpreted as: Copyright 2008-2014 Open Plans : Open Plans contributed to this file EVERY year from 2008 to 2014 These statements aren't true, and the copyright for whatever was done in 2008 hasn't really been transferred anywhere. I think any of the following would be correct: or if you put the same header in all files (those pertaining to TOPP anyway): The only reason to keep Open Plans in there is for sentimental reasons, and then it should probably be prefixed with "Donated by ..." or something like that. |
|
Original comment:
Updating this comment for @arneke (sorry for the short reply earlier I was mobile). My initial short reply was based on the patch accounting for everything needed in a header (a copyright, a date, and an party responsible). The link provided was interesting, and agrees with my experience of file headers. I ask projects to be very careful with copyright as it is the tool we use to enforce our open source licenses. Looking for four things:
Optional:
If the work has been changed from one organization to another (either by the code being donated or by copying code under an compatible open source license) then I expect to see a copyright statement (copyright+date+party) for each organization. I see that this is the point you are concerned about.
I treat this as the file being created in 2008 and last modified in 2014 (by the organization OpenPlans). The work was published each year in that range (as part of a release) so your interpretation is valid. This is when the file was first published by the Open Source Geospatial Foundation. This should become a date range as the file is modified. We have this same situation when including code form other projects. Specifically when reusing code from another project the original file header information should all be there (even you have to add a copyright for your additions into the mix). In this case a "new" GeoWebCache project is being started up ( copyright 2014 OSGeo) based on the "old" GeoWebCache project ( copyright 2008-2014 ). We are updating the header with the new organization name, leaving the original header intact. In the case of GeoWebCache copyright has been transferred (a formal letter was sent between the organizations). For your comment that "Open Plans in there is for sentimental reasons", I think it is more a case of habit based on the far more common case of reusing code between projects. |
|
@jodygarnett It is a shame you're not willing to provide any justification for your statements. It makes it really hard to decide whether you understand how copyright law works, or whether you just want to sweep that GSIP under the rug as quickly as possible. |
|
@arneke what is GSIP? I can think of "GeoServer Improvement Process" but that is probably not what you meant? This issue describes OpenGeo/Boudless donating to OSGeo. The organization was actually OpenPlans (formally known as The Open Planning Project ). |
|
@andrew-dwyer Sorry that what was supposed to be an easy first PR ended up as a discussion of correct copyright notice format. If you'd prefer to work on something a bit more interesting and less contentious, I can deal with making any changes Arne and Jody end deciding are necessary once thy come to an agreement while you grab something from the bug tickets. If you want to see this through I certainly won't complain. |
|
@jodygarnett The example you proposed in GSIP-118 is the basis for the script smithkm wrote. Either way, it doesn't matter where it came from, since you've already reiterated it here. So please stop going on tangents (I worked for TOPP when it became OpenGeo) and focus on the issue: What does the copyright year mean , and why do you think so ? I second the apology to @andrew-dwyer ;) |
|
@jodygarnett Ok, I'm not sure when you updated your comment or when I last loaded the page, but thanks for doing it. So our disagreement is that you emphasise "the work was published each year in that range", and I don't think that publishing matters at all (apart from the very first publication, usually equals copyright year). Only significant contributions / modification warrant a new copyright year. Publishing (i.e. rolling a release) by itself does not extend any rights. Example: With respect to headers you are saying this is the same situation as copying code from other projects, but it is not. Copying text into a file is about two different pieces of code in the same file. You insert the copyright from the source, and you don't change the year unless you go back and copy a newer version. That is a completely different situation from transferring copyright. When you transfer the code is the same as before (no reason to update the year), but the code is now owned by a new entity. The old entity, TOPP, has rescinded all ownership. Therefore it shouldn't be in the header either, because it is of concern to noone. I agree that "Copyright 2007-2015 SOME.ORG" in every file is a valid way of writing copyright headers for a project if you do it from the start of the project and ensure SOME.ORG somehow retains all copyright. Your suggestion to do it at this point doesn't work, in my opinion, because even if you can argue for "2007-2015" there's no single "SOME.ORG" you can put at the end of that line. E.g. there are significant contributions from Boundless that I understand are not covered by the donation. So that forces us to go back to updating just the TOPP / OpenGeo lines on a per-file basis, and as I've argued initially I don't think the copyright year should be changed. This boils down to whether publishing justifies a new copyright year (, and if that were the case then it is unclear whether OSGeo or smithkm / Boundless is the publisher). |
|
Thanks for the discussion @arneke (and sorry if I gave any offence with my initial short reply). I agree that initial file creation is the most significant data point, and we can hold off updating the headers for publishing. The situation of copying code from other files is completely - it is just one we are comfortable with (and happens more often). I agree that with the letter in place we can change TOPP to OSGeo and call this job done (no need to keep TOPP in the copyright). The initial file creation (author and organization) should not be changed. |
|
@smithkm No problems. I'll leave this PR to you. I'll take a look at the list of issues and grab something to work on. |
|
Based on clarification from OSGeo legal you are right @arneke - closing this pull request. |
OpenGeo/Boundless donated copyright of GWC code to OSGeo in 2014.
Update all headers where it can be determined that the code was contributed by OpenGeo or Boundless.
I ran Kevin's ruby script to update the copyright of the matching files. Original discussion here. I'm not sure how Arne's comments on editing the copyright date affect these changes?
@jodygarnett @smithkm @arneke