Submit a paper to JOSS

[gully]

The time has come. muler appears to have reached some threshold level of maturity. It has a feature set that can accomplish the key tasks it was built for. It has months worth of user testing under a range of scenarios, and while it is not perfect, it's pretty good! It is relied upon for projects that are, themselves, destined for publication, and so a citeable reference is due. Also, its API has more-or-less stabilized to a point that it's not likely to have major breaking changes in the future.

So with all that said, let's submit a paper. We will plan to submit to JOSS, because they offer excellent two-reviewer reviews of the actual GitHub-based code. I don't believe it's possible to group muler with its sibling package gollum, but that's OK--- they'll each get their own standalone review.

What we mean by research software

muler is in-scope for JOSS because it meets these criteria:

1. Solves complex modeling problems in a scientific context ✅ It has detail sky fiber modeling built-in, deblazing models, and telluric masking
2. Supports the functioning of research instruments ✅ Makes it easy to use pipeline-reduced data from 3 distinct astronomical instruments
3. Supports the execution of research experiments ✅ Makes it easy to experiment with the order-of-operations of post-processing steps, and their tunable inputs
4. Extracts knowledge from large data sets ✅ Can distill a high-bandwidth echelle spectrum (~57,000 data points), into an industry-standard scalar summary statistic (e.g. equivalent width)

Substantial scholarly effort

muler has taken months of work, with many contributors, commits, issues, pull requests, and considerable on-line and off-line discussions put into its software architecture. It meets these requirements:

• ✅ enables some new research challenges to be addressed (e.g. sky subtraction, telluric masking, deblazing)
• ✅makes addressing research challenges significantly better (e.g., faster, easier, simpler)

Here are the numbers at-a-glance for muler:

• Age of software: 1 year, 2 months
• Number of commits: 254
• Number of authors: 8
• Lines of code: 1888 in the key source code, much more in docs and tutorials
• Whether the software has already been cited in academic papers: papers currently in prep.
• Whether the software is sufficiently useful that it is likely to be cited by your peer group. yes

[gully]

@astrocaroline recommends pursuing this AAS publication and review by JOSS path. This seems like a sweetspot in getting the JOSS review and the AAS audience.

[gully]

Ok, the paper is ready for approval and edits from co-authors.

Our tentative goal is to submit it next Friday, February 25, 2022.

Here is the current draft: muler_JOSS_20220218.pdf

[gully]

This paper is submitted but the review hasn't started.