<div dir="ltr"><div id="gmail-:1k2" class="gmail-Am gmail-aO9 gmail-Al editable gmail-LW-avf gmail-tS-tW gmail-tS-tY" aria-label="Message Body" role="textbox" aria-multiline="true" tabindex="1" style="direction:ltr;min-height:85px">Hi Alex, Hi William,<br><br>Alex chose a specific license that is not really very compatible if you mix it - this is the Non attribution portion. <br><br>However, CC0 is actually one of the most permissive CC licenses. Take a look at the diagram in this page:<br><a href="https://en.wikipedia.org/wiki/Creative_Commons_license">https://en.wikipedia.org/wiki/Creative_Commons_license</a><br><br>And William, CC licenses can be used for software. Here is an example you may have forgotten about: <a href="https://github.com/InstituteforDiseaseModeling/covasim/blob/master/LICENSE">https://github.com/InstituteforDiseaseModeling/covasim/blob/master/LICENSE</a><br><br>However CC0 is much better even from BSD and MIT since it releases from the burden of much paperwork when you reuse it. In large models it may be significant. And we will be seeing larger models in the future.<br><br><br><br>And since Alex brought up a non-commercial license - I want to remind you that those are really a problem. I had at least 2 examples during the pandemic when I needed to use information from papers that used that license and could not get a response after multiple attempts. If taxpayer money funds research that is published under a certain license, the taxpayer should not be restricted from reusing the product of this research. <br><br>There are moral implications here, and like William I do not want to extend the discussion since there is much more work to do on this paper and I want to focus on something we can agree on to reach our deadline in time. Will be happy to discuss more after we get the paper edits ready.<br><br>For now, I suggest that we summarize our dissuasion and if needed edit the paper to a formula we can agree about - I believe that the minimalistic approach in the paper is good enough if we point to this discussion thread for readers that want to read more. <br><br>I am also ok with using this text:<br>"<br>We therefore recommend that models and their associated data should be published under permissive terms. For maximizing reproducibility and integration, we suggest that the most permissive license possible should be chosen. In that regard the CC0 license would be a good choice, effectively waiving interests of the creator in their works and therefore emulating the public domain in jurisdictions where this is necessary. <br>"<br><br>If there are no objections to this sentence, I suggest we use it in the paper and connect it to our discussion. I agree with William, if we want additional discussions we can write a whole paper about it later. Hopefully this compromise is acceptable. <br><br> Jacob<br> <br><br> <br> <br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Sun, Jan 2, 2022 at 8:22 AM William Waites <<a href="mailto:wwaites@ieee.org">wwaites@ieee.org</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">> We therefore recommend that models and their associated data should be published under a Creative Commons license<br>
> which provide a simple, standardized way to give permission to share and use creative work.<br>
<br>
This suggestion is a problem: CC licenses are not intended for<br>
software. CC0 is the only one that works for software. Models<br>
are software.<br>
<br>
> I really feel that being more flexible on the licensing issue is important. I hope that you can agree to my suggestion.<br>
<br>
I agree with this. There is no problem in my view releasing<br>
software under well-established BSD or MIT licenses. They are<br>
pretty much universally compatible and require only attribution.<br>
I do not see a strong argument for saying that people cannot<br>
require attribution. It is, after all, the norm in academia to<br>
the extent that it’s considered misconduct to not do it!<br>
<br>
As I mentioned, I favour the GPL for software licensing myself,<br>
unless contributing to somebody else’s project where they’ve <br>
already chosen a license. But I won’t die on that hill here<br>
because I know that others strongly disagree.<br>
<br>
If we separate out software and data, and do not require CC0,<br>
the ODL is also a reasonable choice for data. <br>
<br>
Flexibility is good. But this is a complicated area. There is<br>
scope for an entire paper on licensing issues. This paper is<br>
already much too long. Perhaps we should trim this section,<br>
sketch out the topic in broad strokes (like I did in my first<br>
paragraph up thread) and leave a detailed exposition to future<br>
work?<br>
<br>
-w<br>
<br>
</blockquote></div>