0

For some research software I have written, I needed some atomic data from NIST. I wanted to put that software onto GitHub (MIT license) but I am not sure if I am allowed to simply also upload the NIST dataset which I need. The NIST copyright page is not too helpful here, it only tells me how to cite the NIST dataset in a paper.

The source code has some comments in it clearly stating the origin of the data, same for a readme file. Would you consider that as enough?

5
  • 1
    In doubt, I'd contact NIST: at the bottom of this page there is a contact email.
    – Massimo Ortolano
    Oct 5 '20 at 18:55
  • How did you obtain the data? Oct 5 '20 at 19:09
  • @MassimoOrtolano good point, I just thought this might be an often encountered problem and a lot of researchers out there have dealt with it
    – Alf
    Oct 5 '20 at 19:50
  • @AzorAhai--hehim you can download a wide variety of datasets directly from their webpage.
    – Alf
    Oct 5 '20 at 19:50
  • @Alf If I (i've never made an account or gotten permission from NIST) could go and download the data right now, there should be no issue with you rehosting it. Oct 5 '20 at 19:56
1

Their copyright page seems fairly clear to me. It even goes into what to do if you create derivative works based on it. That said, best practice would be to include the download as part of the code. Most platforms have the ability to retrieve a url, for example, if the data is available that way. Also, note that Github is not really intended to store datasets, though it will work for that if the dataset is relatively small. It can greatly slow down other work on the rep though.

If you're unable to include the download as part of the code, perhaps include a set of "data download instructions" in the readme.md file. That alone would be more than I'm used to seeing - many times when I've seen the code from published research, it's provided as if you already know how to get all the data yourself.

1
  • In my understanding that statement about derivative works only refers to the non-Standard Reference Data products, i.e. not if I include some Standard Reference Data - datasets - but maybe I'm misunderstanding something here. Anyways, plan was to include the dataset as a separate file (it is not large) and write clear instruction where this is from and how to obtain it.
    – Alf
    Oct 5 '20 at 19:53
1

You can have work under multiple licenses on a single repository

Typically each file would have its own copyright heading, and on your LICENSE or COPYRIGHT file you would state "Project Foobar. All files under src are copyright 2020 by Alf, available under MIT license (see file MIT-License.txt). Contents of dataset folder are made by NIST, published at https://www.nist.gov/alf's-dataset, and available under XYZ terms.".

Do note that in order to combine works under different licenses, they must be compatible. You would not be able to (legally) use, copy or distribute a work which is © by someone else that has all rights reserved, unless you get their permission to do so.

1
  • I'll probably do something along this line, after figuring out of course if the licenses are compatible.
    – Alf
    Oct 6 '20 at 8:05
0

Have you ever looked at NIST Refprop? I think they offer options to incorporate this purchase package into other software. I am not sure the specific properties you are looking for would be included though.

1
  • 1
    No, but if I understand correctly, I need to purchase the NIST Reference Fluid Thermodynamic and Transport Properties Database (REFPROP), and that is something I am not planning to do
    – Alf
    Oct 6 '20 at 8:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.