[General] JOSS paper review #803

Activity

1. tmigot

   assigned

   ocots

   on Jun 28, 2026
2. jbcaillau

   added a commit that references this issue on Jul 27, 2026

   joss: address reviewer Tangi Migot paper review (#803)

   ...

   3b170b7
3. jbcaillau commented on Jul 27, 2026

   

   jbcaillau

   on Jul 27, 2026

   Member

   More actions

   Thank you @tmigot for these constructive remarks, and for the kind words on the paper. All five points are addressed in #813 (draft PR into the joss branch), which now collects our answers to all of your review issues. Point by point below.

   ---

   1. Make the styling/linking of .jl packages uniform ✅

   You were right that this was inconsistent. We have now settled on an explicit convention and applied it throughout the paper:

   every package or tool is hyperlinked at each of its occurrences, and carries its bibliography tag [@key] only at the first occurrence.

   Concretely:

   • Hyperlinks added where they were missing: ForwardDiff.jl and DifferentiationInterface.jl (the two you pointed at), plus NonlinearSolve.jl, NLPModelsIpopt.jl, Knitro, InfiniteOpt.jl, JuMP, ADNLPModels, ExaModels, MadNLP, CUDSS, and every package named in the Julia packages bullet of State of the Field (RobustAndOptimalControl.jl, QuantumControl.jl, Piccolo.jl, DirectTrajectoryOptimization.jl, OptimalControlProblems.jl, CTBenchmarks.jl, …), which were previously bare.
   • Duplicated citation tags removed, keeping each at its first occurrence only: @Patterson2014, @Verschueren2021, @Andersson2019, @arnstrom2022daqp, @Gagnon_*, @pulsipher2022unifying, @QuantumControl_jl, @shin2023accelerating, @SHIN2024110651, @SHIN2021693, @ADNLPModels_jl, @OptimalControlProblems_jl, @Wachter2006, @rackauckas2017differentialequations.
   • QuantumControl.jl is now cited. Checking the bibliography for this point turned up that the QuantumControl_jl entry was present in paper.bib but never used — thanks for prompting the audit.

   The only names deliberately left unlinked are those inside the bold bullet labels of State of the Field (**Legacy tools (COTCOT, HamPath, NutoPy)**, **Direct-method tools (BOCOP, ACADO, GPOPS-II, acados, nosnoc)**); these are enumerations announcing the topic of the bullet, and each tool is hyperlinked in the bullet body.

   2. Add a citation for MINPACK ✅

   Added, and cited at the MINPACK.jl mention in the Indirect methods paragraph:

   Moré, J. J., Garbow, B. S. and Hillstrom, K. E., User Guide for MINPACK-1, Report ANL-80-74, Argonne National Laboratory, 1980. doi:10.2172/6997568

   3. Companion example citation / version ✅

   Good catch — the entry did point at v1.0.0, while v1.0.1 was released on 2026-06-11. Rather than bumping to the v1.0.1 version DOI, we switched the reference to the Zenodo all-versions (concept) DOI:

   10.5281/zenodo.19663649

   which always resolves to the latest release of joss-oc-example and therefore cannot go stale again. The now-misleading version field was dropped from the BibTeX entry.

   4. Add or repeat the Ipopt citation in the "Modelers and solvers" paragraph ✅ (adapted)

   IPOPT is now hyperlinked in that paragraph, so a reader landing there reaches the solver directly.

   We did not repeat the [@Wachter2006] tag there, though: IPOPT is already cited in the Summary, and repeating the tag would contradict the convention adopted in point 1 above — which we apply to every other tool in the paper, including CasADi, GPOPS-II and acados, whose duplicated tags we removed. We hope the hyperlink meets the intent of your comment; happy to reinstate the repeated tag if you would rather have it.

   5. Clarify the exact class of optimal control problems currently supported ✅

   We agree this was missing for a general readership. The paper now has a compact Problem Class section, right after the Summary, giving the Bolza formulation

   $$g(t_0, x(t_0), t_f, x(t_f), v) + \int_{t_0}^{t_f} f^0(t, x(t), u(t), v), \mathrm{d}t \longrightarrow \min$$

   subject to $\dot{x}(t) = f(t, x(t), u(t), v)$, followed by an explicit statement of scope. Answering your checklist directly (each item verified against docs/src/manual-abstract.md):

   Feature	Supported?
   Fixed and free final time	✅ — the time bounds may be fixed or be components of the optimisation variable v, which also covers free initial time and minimum time problems
   Endpoint constraints	✅ — boundary constraints
   Path constraints	✅ — state, control, and mixed (control and state), imposed at all times
   Box constraints on states and controls	✅ — and on the variable; linear range constraints are passed as dedicated bounds to the NLP solver
   Mayer, Lagrange and/or Bolza costs	✅ — Bolza is the general form, Mayer (f⁰ = 0) and Lagrange (g = 0) the special cases; maximisation is handled by minimising the opposite cost
   Autonomous and non-autonomous dynamics	✅ — the dependences are detected by the parser
   Multiphase problems	❌ — single phase only
   DAEs or only explicit ODEs	❌ — explicit ODEs only
   Hybrid / nonsmooth systems	❌

   Constraints may be equalities or one- or two-sided inequalities, linear or nonlinear, and the variable may appear in any of them. The section also makes the optimisation variable v explicit — a finite-dimensional parameter optimised together with the trajectory — which is part of the supported class but had no counterpart in your list.

   ---

   The regenerated PDF is committed on the PR branch by our Compile JOSS paper workflow. Your remaining issue #805 will be addressed in the same PR.

   👍React with 👍1Reacted by Tangi Migot
4. ocots commented on Aug 1, 2026

   

   ocots

   on Aug 1, 2026

   Member

   More actions

   @tmigot Merged in #762.
5. ocots

   closed this as completedon Aug 1, 2026