NEWS.md

In pnd: Parallel Numerical Derivatives, Gradients, Jacobians, and Hessians of Arbitrary Accuracy Order

pnd 0.dev roadmap (2025-XX-XX)

• UX: make the warnings once-per-session; print the first error in runParallel in Grad
• FEATURE: use higher zero tolerances stepx() for higher-order derivatives, update all step... functions using it
• FEATURE: add options for default accuracy (maybe 4?)
• FEATURE: If h is a character in Grad, extract the gradient directly if the order is 2
• FEATURE: suggest parallelisation if f(x) takes more than 0.002 s
• FEATURE: CR, DW, SW, M, K algorithm for arbitrary derivative and accuracy orders
• FEATURE: if any algorithm returns a step size larger than |x| / max(stencil), throw a warning
• FEATURE: Create control or method.args for Grad with automatic step selection
• FEATURE: Arbitrary mixed orders
• MISC: Write the list of controls on the help page of gradstep() explicitly!
• MISC: Check which packages depend on numDeriv and check compatibility with 10 top!
• MISC: Add links to documentation and tutorials onto the GitHub page.
• MISC: Detailed vignette explaining the mathematics behind the functions with full transparency about the choice of parameters
• DEV: ensure that Grad takes all the arguments of GenD and Jacobian, and vice versa
• DEV: Ensure unit-test coverage >90%
• DEV: Check the compatibility between the function and its documentation
• DEV: Check the release with todor::todor_package(), lintr::lint_package(), R CMD check --as-cran, and goodpractice::gp(checks = all_checks()[!grepl("^lintr", all_checks())])

pnd 0.1.1 (2025-09-04)

• Fix: added a simpler and more reliable fall-back option for step.M
• Feature: added plotting methods for step... functions
• Feature: added arbitrary derivation and accuracy order to Curtis--Reid method

pnd 0.1.0 (2025-05-20)

• Feature: original kink-based algorithm for step size selection step.K()
• Feature: added safety shrinking if FUN(x) is finite but FUN(x+h) is not in all SSS routines
• Feature: added S3 printing methods for derivatives and step sizes
• Feature: removed the diagnostics and report arguments; the iteration information is always saved, but not printed
• Feature: added support for max.rel.error for all step-selection methods
• Feature: all step-search methods now return both the truncation and the rounding-error estimate
• Fix: corrected the wrong formula for the plug-in step size
• Fix: now 1x1 Hessians can be computed (why, though, if second derivatives exist?)
• Fix: added the v argument for numDeriv compatibility

pnd 0.0.10 (2025-04-02)

• Fix: GitHub issue #2 -- checkDimensions could not handle character h passed for auto-selection
• Fix: GitHub issue #1 -- function arguments in ... did non propagate properly to step... functions

pnd 0.0.9 (2025-03-10)

• Fix: fixed a regression with the default step size
• Fix: parallelised Hessians in the same manner as gradients
• Feature: compatibility of Hessian() with the arguments for methods "Richardson" and "simple" from numDeriv

pnd 0.0.8 (2025-03-05)

• Fix: sped up CPU core request diagnostics for 1-core operations
• Fix: Using full paths on Macs

pnd 0.0.7 (2025-03-01)

• Fix: removed obsolete environment creation for cluster export
• Fix: changed physical core detection on Macs
• Misc: the package has been added to CRAN, fewer syntax changes are expected

pnd 0.0.6 (2025-01-25)

• Fix: Derivatives of vectorised functions are working. Example: Grad(sin, 1:4)
• Feature: Auto-detecting the number of cores available on multi-core machines to speed up computations
• Feature: Added plug-in step size selection with an estimated f''' with a rule of thumb
• Feature: Auto-detection of parallel type
• Feature: Added zero tolerance to the default step for a fixed step

pnd 0.0.5 (2025-01-14)

• Feature: Extended the step-selection routines to gradients (vector input x)
• Feature: Parallelisation of step selection in all algorithms
• Feature: Mathur's AutoDX algorithm for step size selection step.M()
• Feature: Added Hessian() that supports central differences (for the moment) and arbitrary accuracy
• Feature: Separate Grad() and Jacobian() that call the workhorse, GenD(), for compatibility with numDeriv

pnd 0.0.4 (2024-06-10)

• Feature: Stepleman--Winarsky algorithm for step size selection step.SW()
• Feature: Automated wrapper for step size selection gradstep()
• Improvement: Safe handling of function errors and non-finite returns in step-size selection procedures
• Improvement: Finite-difference coefficients gained attributes: Taylor expansion, coefficient on the largest truncated term, and effective accuracy (useful for custom stencils)
• Improvement: added unit tests for core features

pnd 0.0.3 (2024-06-01)

• Feature: solveVandermonde() to solve ill-conditioned problems that arise in weight calculation
• Feature: Dumontet--Vignes algorithm for step size selection step.DV()
• Feature: Curtis--Reid algorithm for step size selection step.CR() and its modification
• Feature: Different step sizes for the gradient
• Fix: If the user supplies a short custom stencil and requests a high accuracy order, it will provide the best available order and produce a warning
• Fix: The output of Grad() preserves the names of x and FUN(x), which prevents errors in cases where names are required

pnd 0.0.2 (2023-12-06)

• Fix: bug in stencil calculation

pnd 0.0.1 (2023-09-01)

• Initial release
• Computing finite-difference coefficients on arbitrary stencils
• Computing numerical gradients with reasonable default step sizes
• Numerical Jacobians
• Support for mclapply() on *nix systems only

pnd documentation built on Sept. 9, 2025, 5:44 p.m.