[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
Re: [science-iwg] [location-iwg] DOIs for scientific software
|
Hi Amos,
I've nothing to add. Just two comments regarding the DOI.
If the Eclipse Foundation is minting a DOI then it's possible to construct the DOI following own rules - means: your suggestion using a cryptographic hash / checksum from the repository in the DOI should work.
Using the DOI in a URL that refers / resolves to a landing page finally can be used to land either on a specific page set-up for this purposes or on the respective revision page of the repository that uses the .md file to contain information normally found on landing pages - means: your solution should work here, too. However, the implementation of the landing page is something that has to be discussed since it should fit into existing systems and processes in the Eclipse WGs.
Best,
Martin
On 21 Jul 2015 9:04 pm, "Amos Hayes" <
ahayes@xxxxxxxxxxxxxxxx> wrote:
Hi all. I'm a bit of a lurker but wanted to toss a comment in here. Getting academic credit for software, datasets, etc as well as publications is a great goal and one we struggle with too. Being able to unambiguously identify a digital object of any kind with a DOI is great.
But one of the great impacts of DOIs for datasets in particular is that they help others to be able to point to the same dataset and attempt to reproduce observations or augment findings, etc. One thing missing in many cases is a strong assurance that the dataset hasn't changed.
A given release of a software could also have an impact on results. Citing software with a DOI pointing at a repository, assigned version number is a good first step.
But in addition to pointing and trusting the repositories, a cryptographic hash/checksum would allow for some assurances that you were using the exact same tool as the citing author. An Eclipse software (and data? - I refer to other conversations about LocationTech Data) DOI solution should have that as a goal too I think (as Andrew noted when he mentioned including a pointer to commits). And it should ideally support more than just a single repository type or checksum scheme. Git makes this relatively easy due to the nature of git's SHA1 commit hash system. Something including the same for source tarballs and/or binary build checksums would need to happen for subversion, cvs, or other repositories I think.
Github has some info about making code in their system citable. They have it somewhat easy but still, it is an example of something out there:
Andrew, I'm happy to participate in discussions as you see fit. :)
_______________________________________________
location-iwg mailing list
location-iwg@xxxxxxxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://locationtech.org/mailman/listinfo/location-iwg