Join GitHub today
GitHub is home to over 31 million developers working together to host and review code, manage projects, and build software together.Sign up
Save & restore snapshot of ImageJ version with all plugins #17
For scientific reproducibility, it must be possible to publish the exact version of ImageJ used to perform computational analysis, in a form that others can use to repeat said analysis. Unfortunately, the version of ImageJ used is not enough, since it is also paramount to know the version of all plugins used in the analysis, too.
One potential solution would be to hash all the JARs, scripts, macros, etc. -- perhaps the entire ImageJ directory structure. Then index this hash somewhere such that others can restore the same working state from it. We would need to be careful to also describe any non-managed files (i.e., files in the local working copy that aren't part of any ImageJ update site), or at least emit a caveat that these files will not be part of the saved/restored state.
After superficial consideration, I think this is a rather hard, but ultimately doable, problem. But unfortunately we probably will not have time to tackle it until after the initial release of ImageJ 2.0.0.
See this thread on imagej-devel for the full discussion.
Thanks to Michael Doube for raising this issue!
Note that the solution currently available is Make Fiji Package. You would have to post the resultant archive somewhere so that others can download it to repeat your analysis.
The nice thing about this solution is that it includes your entire Fiji installation, including even local-only files. So in theory your analysis should be repeatable by reusing this archive. The downside is that it is less convenient than some built-in program function that gives you a short hash or something which can be published in the paper for all time.