GSAS-II
B. H. Toby and R. B. Von Dreele, Advanced Photon Source, Argonne. Open-source Python with Fortran kernels, successor of GSAS/EXPGUI. Read against the FullProf baseline (L012): only what GSAS-II has and FullProf lacks within the scope, in physics and in workflow, each row saying what FullProf has instead (log L014). GSAS-II sources link to the line in GitHub; FullProf sources are named in manuals/fullprof/.
Sources read. GSAS-II source at commit c13e730 and the GSAS-II-tutorials repository at commit 924e9eb, both cloned from GitHub on 2026-10-09; the help pages are cited from their Markdown sources (MDhelp/docs), from which the in-program HTML help is generated. The published tutorial and help site (advancedphotonsource.github.io/GSAS-II-tutorials) is blocked from this session and is built from the second repository, so its sources were read instead. Every GSAS-II claim below was checked against the cited lines in the clones. The FullProf side is the baseline inventory (data/study/fullprof.yml, 158 entries and its not-found list) and the material in manuals/fullprof/: the 2001 manual, FP_COMMANDS.pdf, the eleven tutorial folders and FP_tests; each candidate gap was searched for in the text of all these PDFs and in the .pcr, .irf, .inf and .bat files before being kept. A capability that both programs have in comparable form is not listed; the ones set aside that way are named in L014. The FullProf side was re-checked against the FullProfAPP manual (version 1.3.3, June 2025, in manuals/fullprof/docs/fullprofapp/): it closes none of the gaps and narrows five (L018).
36 entries.
Data, radiations and patterns
| Capability | What it does | In FullProf | Sources |
|---|---|---|---|
| Binary vendor and HDF5 powder formats | Importers, found as plug-ins at run time, read Bruker .brml and .RAW, Panalytical .xrdml, Rigaku .ras and .txt, CSV, GSAS .fxye (including TOF TIME-MAP), TOPAS .xye, FullProf .dat, powder CIF, MAX IV NeXus HDF5 and MIDAS zarr; most import menus can guess the format from the file. | FullProf reads its own free and instrument formats (Ins=0–14, GSAS files, Socabim .uxd); WinPLOTR reads ASCII formats only, including Panalytical .udf/.xrdml/.csv and Rigaku RINT .txt, and converts them. No binary Bruker .brml/.RAW, HDF5/NeXus or powder-CIF reader appears in the material. | mainmenu.md L101–102 · mainmenu.md L129–134 · G2pwd_HDF5.py L34 · G2pwd_MIDAS.py L49 · FullProf_Manual.pdf, pp. 73–74 · FPS_dec2015_WinPLOTR_TR.pdf, p. 3 · FPS_dec2015_WinPLOTR_TR.pdf, p. 12 |
| Auto Import from a watched directory | A window watches a directory and imports powder files as they appear, filtered by a pattern such as "*June23*.fxye", so a running experiment can be followed in the project. | not found in manuals/fullprof/; FullProf's sequential mode and FullProfAPP's batch import work on files that already exist. | mainmenu.md L140 · FullProf_Manual.pdf, p. 7 · Presentation.pdf, p. 3 |
| Corrections to 2\(\theta\), intensity or weights at import | An .instprm instrument file may carry Python code (CorrectionCode) that is run on import and changes the 2\(\theta\), intensity or weight arrays, e.g. a quadratic 2\(\theta\) calibration correction applied to every pattern from that instrument. | FullProf's Cor=1/2 corrections multiply the calculated intensities by a tabulated or polynomial factor of the scattering variable; there is no correction of the observed x axis or weights at read time. | mainmenu.md L124–127 · others.md L125–145 · FullProf_Manual.pdf, p. 71 · FullProf_Manual.pdf, pp. 130–131 |
| Phase import from many structure formats | Phases are read from CIF, SHELX .ins/.res, GSAS .EXP, JANA .m50, PDB, ICDD .str, RMCProfile .rmc6f, RRUFF DIF, XYZ and other GSAS-II projects; the coordinates of an existing phase can later be reset from any importable file (Reimport atoms). | WinPLOTR and EdPCR convert CIF and SHELX .ins to .pcr; FullProfAPP fetches CIFs from the COD. No other structure formats are named. | mainmenu.md L111–120 · G2phase_RRUFF.py L19 · phaseatoms.md L52 · FPS_dec2015_WinPLOTR_TR.pdf, p. 9 · Tutorial-Y2O3.pdf, pp. 13–17 · Exercise-EgyptMakeup.pdf, p. 1 |
The crystal structure
| Capability | What it does | In FullProf | Sources |
|---|---|---|---|
| Hydrogen atoms placed from geometry | Calc H atoms inserts H atoms at standard positions on selected atoms from their bonding pattern; Update H atoms resets them to the bonding geometry after a refinement. | not found in manuals/fullprof/; hydrogen atoms are typed in by hand or carried by rigid bodies and satellite groups (e.g. methyl groups), and FIX_SPC H fixes them. | phaseatoms.md L41 · phaseatoms.md L48 · FullProf-Tutorial-RigidBody.pdf, p. 12 · FP_COMMANDS.pdf, p. 3 |
Peak shapes, instrument and peak positions
| Capability | What it does | In FullProf | Sources |
|---|---|---|---|
| Pink-beam profile | A constant-wavelength histogram type 'B' for pink (wide-bandpass) beams: back-to-back exponentials convolved with a pseudo-Voigt, with α = α0 + α1·tanθ and β = β0 + β1·tanθ for X-rays (sinθ for neutrons), plus U, V, W, X, Y, Z (Von Dreele, Clarke & Walsh 2021; Von Dreele 2024). | FullProf uses back-to-back exponentials ⊗ pseudo-Voigt only for TOF (Npr=9, 10); its CW profiles are Gaussian, Lorentzian, pseudo-Voigt, Pearson VII, TCH with FCJ asymmetry, split and numerical. 'Pink beam' is not found in manuals/fullprof/. | GSASIIpwdGUI.py L3070–3075 · GSASIIpwd.py L934–937 · GSASIIpwd.py L1426–1434 · GSASIIstrMath.py L4093–4094 · FullProf_Manual.pdf, p. 69 · FullProf_Manual.pdf, pp. 111–112 |
| Instrument profile from fundamental parameters | For a CW (usually Bragg–Brentano) instrument the NIST FPA code computes peaks from Topas-style inputs (divergence, Soller angles, goniometer radius, filament, sample and receiving-slit lengths, absorption, thickness, tube tails, PSD range, incident-beam monochromator) and a source spectrum, then fits U, V, W, X, Y and SH/L to them and saves an instrument parameter file. FPA is not used inside the Rietveld fit. | FullProf's resolution file (IRF) is obtained by refining a standard such as LaB6; the only geometric term is the Finger–Cox–Jephcoat axial divergence (S_L, D_L). Fundamental-parameter peak shapes are on the baseline's not-found list. | mainmenu.md L142–144 · others.md L82–122 · FullProf_Manual.pdf, pp. 131–132 · FullProf_Manual.pdf, pp. 112–113 |
Background
| Capability | What it does | In FullProf | Sources |
|---|---|---|---|
| Background peaks | Any number of pseudo-Voigt background peaks, each with refinable position, integrated intensity, Gaussian width and Lorentzian width, added to a smooth function; meant for broad lumps such as Kapton or container scattering. | FullProf models a parasitic peak as an extra phase with fixed relative intensities read from an .hkl file and a refined scale (Jbt=3, Irf=2), or absorbs humps with Debye-like terms or a background file; no free-standing refinable background peak. | powderbkg.md L31–35 · HoBaCuO-HT.pcr L216–225 · FullProf_Manual.pdf, pp. 45–46 |
| Background function fitted to fixed points, with auto-background | Fixed points are placed with the mouse or generated by pybaselines (arpls or iarpls, with a log(λ) smoothing slider); Fit Background then refines the chosen function (e.g. Chebyshev with a few terms) to the points before any Rietveld cycle. The auto-background can instead be saved as a fixed background histogram subtracted point by point. Gap narrowed to: fitting functions other than Chebyshev to the points, and an automatic background used as a fixed background. (L018) |
WinPLOTR picks background points by hand or automatically (about √n points, "not recommended for advanced analysis") and saves them to paste into the .pcr, where they become a linear-interpolation or spline background; FullProfAPP extracts background points (method not stated). The points are not used to fit a polynomial or Chebyshev series. FullProfAPP finds background points automatically (asymmetric least squares, two sliders), lets them be edited on the plot and fits a Chebyshev series to them as starting values. | powderbkg.md L92–118 · powderbkg.md L63 · FPS_dec2015_WinPLOTR_TR.pdf, pp. 38–40 · FullProf_Manual.pdf, p. 78 · Exercise-LiFePO4.pdf, p. 2 · index.html L242–246 · index.html L137–146 · calculators.html L131–149 |
| Background in powers of Q, and inverse or logarithmic interpolation | Besides Chebyshev and cosine series, the background function can be a power series in Q\(^{2}\) or in Q\(^{-2}\), or an interpolation through points spaced equally in 1/x or in log x (inv interpolate, log interpolate), with 1 to 36 terms. | FullProf offers a polynomial in (T/Bkpos − 1), Chebyshev (up to 24 terms), cosine Fourier series, linear interpolation and cubic splines through user points; no Q-power series and no inverse or logarithmic point spacing. | GSASIIpwdGUI.py L2280 · GSASIIpwd.py L1030 · powderbkg.md L9–18 · FullProf_Manual.pdf, pp. 45–46 · pbso4_cheb.pcr L26 |
Size and strain broadening
| Capability | What it does | In FullProf | Sources |
|---|---|---|---|
| Gauss–Lorentz mixing of anisotropic size broadening | Isotropic, uniaxial and ellipsoidal size broadening each carry an LGmix parameter that splits the size width between Lorentzian and Gaussian parts (1 = pure Lorentzian), as microstrain does: \(\gamma\) = S·mx_size + M·mx_strain, \(\sigma\)² = (S(1−mx_size))² + (M(1−mx_strain))². | FullProf has a Gaussian isotropic size term (IG/GauSiz) and a mixing parameter ξ for anisotropic strain, but its anisotropic size models are Lorentzian only ("the anisotropic size is supposed to contribute to the Lorentzian"). | phasedata.md L29–38 · GSASIIstrMath.py L3330–3332 · FullProf_Manual.pdf, p. 23 · FullProf_Manual.pdf, p. 21 |
Intensity corrections
| Capability | What it does | In FullProf | Sources |
|---|---|---|---|
| Spherical-harmonic preferred orientation | Per phase and histogram, a cylindrically symmetric spherical-harmonic orientation distribution of chosen even order, constrained to the crystal symmetry, as an alternative to March–Dollase; the fitted distribution can be checked as a ψ-scan or inverse pole figure in multiples of random distribution. | FullProf has the exponential (Nor=0) and modified March–Dollase (Nor=1) functions about one direction, with an untextured fraction; texture beyond these is on the baseline's not-found list. | phasedata.md L64–77 · GSASIIstrMath.py L3010–3036 · GSASIIstrMath.py L3183–3190 · FullProf_Manual.pdf, pp. 54–55 |
| Texture analysis with generalised spherical harmonics | A phase-wide orientation distribution function (Bunge; Von Dreele 1997) of even order up to 34, with sample symmetry cylindrical, 2/m, mmm or none and refinable sample-orientation zeros, applied to all histograms of the phase recorded at different sample orientations; it takes precedence over the per-histogram corrections. The texture index J is reported, pole figures, inverse pole figures and 3D pole distributions are drawn, and a texture restraint is offered. | not found in manuals/fullprof/; the QPA tutorial's paper on textured alloys obtained ODFs from pole figures measured and computed outside FullProf, then compared them with March–Dollase in the Rietveld fit. | phasetexture.md L7–13 · phasetexture.md L112–128 · GSASIIstrMath.py L3184–3185 · GSASIIrestrGUI.py L2540–2544 · Quantitative-prefOrient.pdf, pp. 2–3 |
| Extinction for constant-wavelength powder data | One extinction coefficient per phase and histogram for CW X-ray and neutron as well as TOF powder data, combining Bragg and Laue components weighted by sin\(^{2}\)\(\theta\) and cos\(^{2}\)\(\theta\) (Sabine-type), with analytic derivatives. | FullProf has one powder extinction parameter for TOF data only, and a SHELX-type parameter for single-crystal and integrated data; CW powder extinction is on the baseline's not-found list. | GSASIIstrMath.py L3121–3146 · GSASIIstrMath.py L3194–3196 · phasedata.md L79 · FullProf_Manual.pdf, p. 108 · FullProf_Manual.pdf, p. 110 · FullProf_Manual.pdf, p. 63 |
Le Bail and Pawley fits
| Capability | What it does | In FullProf | Sources |
|---|---|---|---|
| Pawley refinement | Reflection intensities are least-squares parameters refined with the cell and profile, between a Pawley dmin and dmax, with an optional penalty weight on negative intensities; reflections are created, estimated from peak heights, and negative ones reset and unflagged. Intensities thus get standard uncertainties and enter the covariance matrix; the histogram scale must be held. | FullProf implements Le Bail extraction (Jbt=2: intensities iterated from the Rietveld partition, not refined); the manual names Pawley's method only as a relative of profile matching. | phasegeneral.md L75–78 · phasepawley.md L12–14 · FullProf_Manual.pdf, pp. 14–15 · FullProf_Manual.pdf, p. 89 |
Constraints, restraints, rigid bodies
| Capability | What it does | In FullProf | Sources |
|---|---|---|---|
| Linear constraint equations and New Var parameters | Four user constraint types on named parameters: Hold, Equivalence (fixed ratios), Constraint equation (Σ mᵢpᵢ = constant, removing one degree of freedom from any number of parameters, e.g. three cations on one site summing to 1 with two free) and New Var (a new refined parameter defined as a weighted sum). Symmetry-generated constraints are shown separately; the processed relations, errors and warnings can be displayed; constraints export to and import from .constr text files; Make atoms equivalent sets up a shared site in one step. | FullProf ties parameters by sharing a refinement code with multipliers, or by EQUAL/PEQU keywords; each group has one free shift, so a sum over three or more parameters with more than one free direction cannot be imposed exactly; FullProf offers soft linear restraints (NLI) for that. | commontreeitems.md L108–120 · commontreeitems.md L131–179 · commontreeitems.md L185–187 · FullProf_Manual.pdf, p. 19 · FP_COMMANDS.pdf, p. 3 · FullProf-Tutorial-RigidBody.pdf, pp. 4–5 |
| Histogram wildcards in sequential constraints | In sequential refinement a constraint can name a histogram as * (p:*:name) so it applies to every histogram in turn; three modes decide whether numbered constraints are converted to *, ignored, or used as written. | not found in manuals/fullprof/; a FullProf sequential run reuses the single .pcr, whose codes apply to whichever pattern is being refined. | commontreeitems.md L193–199 · FPS_dec2015_WinPLOTR_TR.pdf, pp. 18–19 |
| General restraints as Python expressions | A restraint on any quantity computed by a user-supplied Python expression of GSAS-II parameters, with target and s.u.; conditional (if) logic makes one-sided minimum or maximum restraints. Each restraint type has its own weight factor and on/off switch. | FullProf restrains distances, angles and linear combinations of refined parameters (NLI), all two-sided, with the restraint term weighted by the current reduced \(\chi^2\). | commontreeitems.md L239 · commontreeitems.md L213–214 · FullProf_Manual.pdf, p. 42 · FullProf_Manual.pdf, pp. 123–125 · FullProf-Tutorial-RigidBody.pdf, pp. 4–5 |
| Plane and chiral-volume restraints | Restraints that keep a chosen set of atoms nearly coplanar, and restraints on the chiral volume of an atom, available for every phase type (torsion and Ramachandran restraints only for macromolecular phases). | FullProf has distance, angle and linear restraints only; plane and chiral restraints are not found in manuals/fullprof/. | GSASIIrestrGUI.py L2503–2511 · commontreeitems.md L234–237 · FullProf_Manual.pdf, pp. 123–125 |
| Restraints imported from Mogul | Bond and angle restraint targets and s.u.s are read from the .csv written by the CCDC Mercury/Mogul geometry check (or a file in that layout), i.e. from statistics of the Cambridge Structural Database. | FullProf writes candidate restraint lines from its own distance calculation (DCONSTRn.hlp, CFML_Restraints.tpcr) for pasting into the .pcr; Mogul is not found in manuals/fullprof/. | commontreeitems.md L222–225 · commontreeitems.md L242–263 · FullProf_Manual.pdf, p. 124 · CFML_Restraints_1.tpcr L1–11 |
The least-squares engine
| Capability | What it does | In FullProf | Sources |
|---|---|---|---|
| Levenberg–Marquardt least squares with SVD | The default minimiser builds the Hessian from analytic derivatives histogram by histogram, inverts it by singular-value decomposition with a zero tolerance (default \(10^{-6}\)), so near-singular directions are dropped and the parameters causing them are named, and up-weights the Hessian diagonal (Marquardt λ) when a cycle fails to lower \(\chi^2\). Alternatives: SVD without Marquardt, MINPACK leastsq with analytic or numerical derivatives (the latter to check code). Stops on Min ΔM/M (default 0.001) or Max cycles. | FullProf solves the Gauss–Newton normal equations, damping shifts with class relaxation factors and code multipliers; a singular matrix stops the run and the manual advises restarting with scale factors only; Dum=2 stops when \(\chi^2\) rises. No SVD or Marquardt step is mentioned. | commontreeitems.md L36–50 · GSASIImath.py L164–246 · GSASIIstrMain.py L405–436 · FullProf_Manual.pdf, pp. 42–43 · FullProf_Manual.pdf, p. 77 · FullProf_Manual.pdf, p. 16 · FullProf_Manual.pdf, p. 68 |
| Parameter impact | Ranks the parameters not being refined by the improvement in reduced \(\chi^2\) each would give if refined (Toby, J. Appl. Cryst. 57, 175–180, 2024); also callable from scripts as ComputeWorstFit. | not found in manuals/fullprof/; FullProf documents a fixed refinement order (scale; zero, background, cell; positions; shape; occupancies; displacements; microstructure). | mainmenu.md L85 · GSASIIscriptable.rst L317 · FullProf_Manual.pdf, p. 14 |
| Any expression with its s.u. from the full covariance | Evaluate Expression and s.u. computes a user expression of several parameters (e.g. the total amount of an element spread over sites) with a standard uncertainty propagated through the covariance matrix, so correlations are included. | FullProf prints s.u.s of refined parameters and of its own derived quantities (distances, weight fractions); in the symmetry-mode output, atom-position s.u.s are propagated "neglecting the correlation between the refined amplitudes". No user expressions. | mainmenu.md L87 · FullProf_Manual.pdf, p. 43 · FullProf-Tutorial-Symmetry-Modes.pdf, p. 4 |
| Normal probability plot | Error Analysis draws a normal probability plot of the weighted residuals within the limits; the slope over the central ~68 % of points (−1 to 1) is the GOF of the well-fitting points and the intercept shows bias. | FullProf reports Durbin–Watson d, the Bérar factor for serially correlated residuals, Akaike and Schwarz criteria and an expected-R analysis (Ana=1); a normal probability plot is not found in manuals/fullprof/. | powderparent.md L21 · FullProf_Manual.pdf, p. 45 · FullProf_Manual.pdf, pp. 61–62 · LaMnO3ref.out L796–799 |
Workflow: running, sequences, programs around it, the live display
| Capability | What it does | In FullProf | Sources |
|---|---|---|---|
| One project file with a data tree | Everything — data, instrument and sample parameters, phases, constraints, restraints, rigid bodies, results, covariance, sequential tables, notebook — lives in one .gpx file shown as a data tree; any number of phases and histograms, limited by memory; each save during refinement keeps a numbered backup <project>.bak<i>.gpx that can be reopened. Gap narrowed to: data, model, constraints, results and covariance in one file; the model stays in .pcr files. (L018) |
FullProf keeps model and flags in a .pcr text file, data in separate files and results in .out, .sum, .prf and other files; the .pcr is rewritten after a run (Pcr=1) or written to CODFIL.new (Pcr=2); up to 16 phases. FullProfAPP saves the interface state in a .fpapp file, shows the working folder as a tree, and writes each run to a new numbered folder. |
datatree.md L5 · datatree.md L34–36 · mainmenu.md L14–15 · FullProf_Manual.pdf, pp. 8–9 · FullProf_Manual.pdf, p. 72 · FullProf_Manual.pdf, p. 87 · index.html L130–153 · index.html L437–446 · index.html L565 |
| Python scripting API | GSASIIscriptable reads and writes the same .gpx files as the GUI: add histograms (real or simulated), phases, images and PDFs; set values and flags; run lists of refinement steps; add Hold, Equivalence, Equation and New Var constraints; read covariance; rank parameters. It runs without the GUI packages, e.g. on a cluster, but does not cover every GUI feature. Gap narrowed to: running refinements from a script, reading the covariance, adding constraint equations; the modules are presented as the app's utilities, not a supported interface. (L018) |
FullProf is driven by the .pcr and the command line (fp2k code data log), WinPLOTR command files for data handling, and the FullProfAPP GUI for batches; no programming interface to FullProf is described. FullProfAPP documents Python modules that read and edit .pcr files (flags, background, regions, phases, VARY/FIX, limits, restraints) and parse .sum, .prf and .mic results. |
GSASIIscriptable.rst L12–19 · GSASIIscriptable.rst L54–58 · GSASIIscriptable.rst L248–322 · FullProf_Manual.pdf, pp. 6–7 · FPS_dec2015_WinPLOTR_TR.pdf, pp. 42–44 · Presentation.pdf, p. 3 · index.html L100 · index.html L172 · app_io.html L1180–1187 |
| Pseudo-variables and parametric fits of sequential results | Sequential results form a table (one row per histogram, one column per parameter); clicking a column plots it with error bars against histogram or any column such as temperature; pseudo-variables (formulas, distances, angles) add derived columns, and parametric equations (e.g. a thermal-expansion law) are fitted to columns with errors propagated from each refinement's covariance; rows reopen their fit and covariance. | FullProf's sequential mode stores the refined parameters of each pattern in a .seq table that WinPLOTR plots; derived quantities and fits of equations to the series are not found in manuals/fullprof/. | sequential.md L19–55 · sequential.md L66–70 · FullProf_Manual.pdf, p. 7 · FPS_dec2015_WinPLOTR_TR.pdf, pp. 18–22 |
| Accept or reject a refinement | After a fit the user sees a table of how each parameter changed and either loads the results or rejects them, leaving the project unchanged so that flags or settings can be changed and the fit tried again. | FullProf overwrites the .pcr with the refined values (Pcr=1) or writes them to CODFIL.new (Pcr=2), leaving the choice to file handling; wfp2k only asks whether to continue or stop. | mainmenu.md L78–80 · FullProf_Manual.pdf, p. 72 · FullProf_Manual.pdf, p. 16 |
| Project notebook with logged parameters | A time-stamped log of refinements, Fourier maps and peak fits, with user comments, filters and a plot of Rw, reduced \(\chi^2\) or chosen logged parameters against refinement number; saved as <project>-notebook.txt. | FullProf writes per-run summaries (.sum, .rpa appended run after run) and comment headers with \(\chi^2\) and R_Bragg in the rewritten .pcr; no log plotted across refinements. | commontreeitems.md L9–20 · mainmenu.md L71–72 · FullProf_Manual.pdf, p. 9 · FullProf_Manual.pdf, pp. 65–66 |
| Limits and excluded regions set on the plot | Data limits and excluded regions are drawn as vertical lines on the pattern and set by dragging, by menu clicks or by pressing e at each end; limits can also be dragged in Q or d. | FullProf takes excluded regions as low/high pairs typed on line 11 of the .pcr (WinPLOTR draws them); setting them graphically is not described. | powderlimits.md L11–14 · powderlimits.md L52–59 · FullProf_Manual.pdf, p. 78 · FPS_dec2015_WinPLOTR_TR.pdf, p. 13 |
| Updates and regression through git | The Help menu checks for and installs newer versions and can regress to any earlier version, e.g. to reproduce an old result or avoid a new bug. Software distribution rather than refinement; kept because the scope counts workflow. |
not found in manuals/fullprof/; FullProf and FullProfAPP are downloaded as installers from the ILL site. | mainmenu.md L194–198 · Presentation.pdf, p. 5 |
Output files
| Capability | What it does | In FullProf | Sources |
|---|---|---|---|
| Fit-diagnostic plot modes | Key commands on the Rietveld plot switch the difference curve to (obs − calc)/\(\sigma\) (w), superimpose a cumulative \(\chi^2\) curve showing where the misfit accumulates (X), add magnified regions (a), subtract the background (b), plot phase partials (p), and show d or Q axes; hovering over a tick mark shows its hkl, and right-click labels persist in the project. | WinPLOTR plots the .prf with linear, log, √ and I·Q\(^{2}\) scales, x in 2\(\theta\), TOF, energy, Q or d, and hkl on click; weighted differences and cumulative \(\chi^2\) are not found in manuals/fullprof/. | powder.md L85–106 · powder.md L75–78 · FPS_dec2015_WinPLOTR_TR.pdf, pp. 3–4 · FPS_dec2015_WinPLOTR_TR.pdf, p. 7 |
| Interactive correlation and shift displays | The Covariance item draws the correlation matrix as a colour map (−1 to 1) with parameter names and values on hover, and bar charts of shift/s.u. over the last refinement or the last cycle. | FullProf writes the correlation matrix to the .out (Mat=1) or prints the matrix diagonal each cycle (Mat=2); no plot. | commontreeitems.md L83–97 · FullProf_Manual.pdf, p. 72 |
| Publication plots | The P button opens an editable copy of the Rietveld plot (colours, line widths, limits, text) exported as bitmap, PDF, PostScript, SVG, Grace, Igor Pro, Origin or CSV. Gap narrowed to: named vector and program formats (PDF, SVG, Grace, Igor, Origin) and an editable publication copy of the Rietveld plot. (L018) |
WinPLOTR saves graphics as bitmap or PostScript, or in its own .wpl/.pgf formats. FullProfAPP changes graph properties and saves plots from a context menu, and exports graph data as XY text. | powder.md L162–166 · FPS_dec2015_WinPLOTR_TR.pdf, p. 4 · FPS_dec2015_WinPLOTR_TR.pdf, p. 42 · index.html L364 · index.html L312 |
| Plots of the refined size, strain and orientation models | The Phase/Data tab draws 3D surfaces of the crystallite-size and microstrain models, and the preferred-orientation model as a ψ-scan or an inverse pole figure. | FullProf lists apparent size and strain per reflection in the .mic file and in the .out; plots of the models are not found in manuals/fullprof/. | phasedata.md L88 · FullProf_Manual.pdf, pp. 11–12 · Microstructural_effects_FP.pdf, p. 8 |
| Sequential results as multi-block CIF and CSV | A sequential refinement exports as one CIF with a block per histogram (or per phase and histogram), as Cinema: Debye-Scherrer files, or as CSV of the whole table. Gap narrowed to: a multi-block CIF of the series. (L018) |
FullProf writes a CIF per run and sequential summaries (.rpa/.sav, .seq); a combined CIF for a series is not found in manuals/fullprof/. FullProfAPP exports all refined parameters of a sequential run from its Summary tab as XY text columns. | sequential.md L56–62 · FullProf_Manual.pdf, p. 9 · FullProf_Manual.pdf, p. 72 · PPH3-sls.cif L1–6 · index.html L308–313 · index.html L193–194 · index.html L275 |
Kept for later
- Magnetic structures. Magnetic phases with moment parameters AMx, AMy, AMz; magnetic subgroups from Bilbao k-SUBGROUPSMAG, stepped through on the powder plot; k-vector search; five Magnetic tutorials. GSASIIobj.py L629 · powdercells.md L47–52 · powder.md L155–160
- Modulated structures and superspace. (3+1)-dimensional incommensurate structures with modulations of site fraction, position, displacement and moment (Fourier, crenel, zigzag/block) in the Wave Data tab. phasewave.md L5 · GSASIIobj.py L699–713
- Stacking faults. The Layers tab drives a modified DIFFaX for simulations and parameter sequences; fitting a pattern is "not available yet". phaselayers.md L5–15
- Indexing. Peak search and fit, autoindexing on chosen Bravais lattices with M20/X20, cell refinement, space-group trials, SUBGROUPS and cell-symmetry search; peaks exportable to PreDict. powdercells.md L16–25 · powderpeaks.md L75–87
- Structure solution. Charge flipping (also 4D) from single-crystal or Pawley/Le Bail-extracted intensities, Monte Carlo/simulated annealing with rigid bodies, Fourier and Patterson maps with peak search, maximum-entropy maps via Dysnomia. phasegeneral.md L18–22 · phasemcsa.md L5 · phasedysnomia.md L5
- 2D image integration. Calibration, masks, dark/background/gain corrections and integration of area-detector images to powder patterns; polarisation calibration, 2D strain and texture from images. image.md L35–40 · image.md L20–22
- PDF / total scattering. G(r) computed from a powder histogram with an optimiser for background and Ruland width; modelling through PDFfit2, RMCProfile and fullrmc. pairdistribution.md L5–6 · phaseRMC.md L5–6
Out of scope
- Single-crystal data. HKLF histograms (CW X-ray, CW and TOF neutron), refinement on F\(^{2}\) or |F|, extinction models, merohedral twins, Flack parameter, joint single-crystal and powder fits. singlecrystal.md L5 · objvarorg.rst L1053–1058
- 3D electron diffraction. Single-crystal histogram type SEC for CW electrons (micro-ED) with a microED extinction choice. objvarorg.rst L1053–1058 · GSASIIddataGUI.py L599–601
- Small-angle scattering. SASD histograms with size distributions, particle fits, pair-distance and bead models. smallanglescattering.md L85–90
- Reflectometry. REFD histograms with layered models fitted by four methods. reflectometry.md L85–91
- Macromolecules. PDB import, residue rigid bodies, torsion and Ramachandran restraints, Babinet solvent, errat protein-quality check; judged out as protein crystallography rather than powder structure refinement of the kind in scope. phasegeneral.md L49–50 · phasedata.md L81
- Atom deformation (charge density). Orbital softness, populations and spherical-harmonic deformation parameters per atom; a charge-density refinement, not routine powder work. GSASIIobj.py L630–637
- Cluster and outlier analysis. Unsupervised grouping of PWDR or PDF datasets with scipy and scikit-learn; it acts on raw data, not on refinement results. cluster.md L5–9 · mainmenu.md L89
Sources not reached
- The published GSAS-II sites (tutorials, help, readthedocs) were not opened; their sources in the two clones were read instead.
- Neither program was run: GUI behaviour is taken from the help pages and code, not observed.
- GSAS-II's third CW profile type 'A' (back-to-back exponentials with FCJ asymmetry, getExpFCJVoigt3) is in the code but undocumented in the help, so it was not listed.
- Papers cited by GSAS-II (Toby 2024; Von Dreele 1997, 2024; Von Dreele, Clarke & Walsh 2021; Mendenhall, Mullen & Cline 2015) were not opened.