[Vp-reproduce-subgroup] "Models are not consistently licensed"

[Jacob Barhak]

Hi Alex, Hi William,

Alex chose a specific license that is not really very compatible if you mix
it - this is the Non attribution portion.

However, CC0 is actually one of the most permissive CC licenses. Take a
look at the diagram in this page:
https://en.wikipedia.org/wiki/Creative_Commons_license

And William, CC licenses can be used for software. Here is an example you
may have forgotten about:
https://github.com/InstituteforDiseaseModeling/covasim/blob/master/LICENSE

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.



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.

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.

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.

I am also ok with using this text:
"
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.
"

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.

            Jacob





On Sun, Jan 2, 2022 at 8:22 AM William Waites <wwaites at ieee.org> wrote:

> > We therefore recommend that models and their associated data should be
> published under a Creative Commons license
> > which provide a simple, standardized way to give permission to share and
> use creative work.
>
> This suggestion is a problem: CC licenses are not intended for
> software. CC0 is the only one that works for software. Models
> are software.
>
> > I really feel that being more flexible on the licensing issue is
> important. I hope that you can agree to my suggestion.
>
> I agree with this. There is no problem in my view releasing
> software under well-established BSD or MIT licenses. They are
> pretty much universally compatible and require only attribution.
> I do not see a strong argument for saying that people cannot
> require attribution. It is, after all, the norm in academia to
> the extent that it’s considered misconduct to not do it!
>
> As I mentioned, I favour the GPL for software licensing myself,
> unless contributing to somebody else’s project where they’ve
> already chosen a license. But I won’t die on that hill here
> because I know that others strongly disagree.
>
> If we separate out software and data, and do not require CC0,
> the ODL is also a reasonable choice for data.
>
> Flexibility is good. But this is a complicated area. There is
> scope for an entire paper on licensing issues. This paper is
> already much too long. Perhaps we should trim this section,
> sketch out the topic in broad strokes (like I did in my first
> paragraph up thread) and leave a detailed exposition to future
> work?
>
> -w
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.simtk.org/pipermail/vp-reproduce-subgroup/attachments/20220102/a492973a/attachment.html>