Software Paper Revision

[cortner]

I'm collecting here some thoughts for potential changes in resonse to the referee reports and/or discussions amongst ourselves:

• Double check that no documentation pages give 404 errors before final resubmission.
• Finalize Si section : briefly explain choice of smoothness priors, demonstrate smoothness of potentials ...
• One referee points out that referring to Julia artifacts is confusing. Maybe we should rename ACE1pack.artifact -> ACE1pack.example_dataset or something similar. Could even be an alias instead of a rename. (+ revive this in the first place as it got accidentally removed). Suggestion: combine the artifact + read_xyz into read_example_dataset -> cf WIP: convenient access to example datasets #149
• Molly tutorial ( @tjjarvinen )
• Should the inner cutoff be removed from the software paper entirely? We now never use it I believe? Do any of the examples use it? -> leave as is ...
• Is it ok to show only the simple Agnesi(2,2) transform as a prototype but use more general variants for all of the examples? -> rephrase as "illustrative" in the article
• Rename ACE1pack -> ACEpotentials
• Finalize H2O section
• mention alternative smoothness priors in methods section. (JAMES)
• Tutorials say that lammps export is broken. Is this now fixed? Can those lines be removed?
• TiAl-basis tutorial is broken due to missing row_info in ACEfit.jl
• ACE1pack interface tutorial is also broken and out of date. What will we do with that?

From previous list of TODOs

• rewrite "intro to ACE models" once the manuscript is done and base it on that.
• link to software paper as intro to ACE
• need a page in the docs on how to prepare data to work with ACE1pack Documentation on Preparing Data #95

ASAP AFTER SUBMIT REVISION

• Make all examples available through ACEworkflows. (unless we want to keep them in the ACE1pack repo)
• the TiAl rdf example in the dataanalysis tutorial is incredibly boring, do we have a nicer dataset to add to the artifacts? -> add a small H2O dataset to the artifacts
• Add a function load_ace_potential and save_ace_potential
• Coordinate these steps with ACEmd.jl

URGENT FUTURE

• for future compatibility consider switching ALL models to splines from the start.
• Rename ACE1Model -> ACEpot or something along those lines... -> maybe this is not urgent, can consider for later.
• Improve documentation of transforms