Errata

Updated 2026-11-04. Documentation

Faults found in released versions of RadMap. Each entry says which versions it affects, whether it can change results and under which conditions, and the version that fixes it. Every PDF report names the RadMap version that produced it, so you can tell whether a result needs review.

This page is updated whenever a fault is found, independently of releases. To report a problem, use Report a problem.

D-0001: Sources in a moved parent, or rotated tube segments, sampled in the wrong place

RadMap tells Geant4 to sample decay positions in a box around the source solid and keep those inside it. The box's centre was the solid's position relative to its parent, not in the world, and its offset from the solid's origin was not turned with the solid. So for a source in a moved parent, or a rotated segment whose bounding box is off its origin, the box sat away from the solid. Results of such projects made with 1.x should be recalculated with 2.0.0.

D-0002: Sampling box of a curved source fitted to its drawing mesh, slightly too small

RadMap tells Geant4 to sample decay positions in a box enclosing the source solid and keep those inside it. The box was fitted to the corners of the mesh used to draw the solid, and a curved surface bulges out beyond the mesh between its corners, so the box stopped short of the solid there. The box is now computed exactly from the solid's dimensions.

D-0003: Effective dose used ICRP 116's coefficients without their 2015 corrections

ICRP corrected 199 of the photon effective dose coefficients in Table A.1 of Publication 116 in 2015 (Corrigenda, Annals of the ICRP 44(1)). RadMap's tables were transcribed from the original 2010 table and carried the uncorrected values, although the corrections were published before RadMap's first release. RadMap now uses the corrected table, and its data source reads ICRP 116 (2010, corrected 2015).

D-0004: Isotopes offered whose ground state has no decay data emitted nothing

The isotope list offered every isotope Geant4 has a decay file for. For some, the file describes only an excited state of the nucleus, and Geant4 can then not decay the ground state a source starts in. In the decay data 1.1.0 used (Geant4's RadioactiveDecay 4.3.2) that held for the nine isotopes above; in the data 2.0.0 uses it holds for 39. The isotope list now leaves out every isotope whose ground state has no decay data, and a project that holds one is refused before the simulation starts, with a message naming the source.

D-0005: Deep dose maps read the radiation where it entered them, not their mean

A dose map is meant to read the dose rate averaged over its volume. It counted the photons entering it through any face, each divided by the area of the box seen from its direction. That is the mean over the box only where the radiation is the same throughout it, as for a thin plate. Where the radiation changes across a deep box, the count follows the faces the photons enter by, which depend on how the box is turned, so the same box in the same place read differently in different orientations. RadMap now adds up the length of every photon's path inside the box and divides by its volume, which is the mean over the volume by definition, in any orientation and wherever the photons come from.

D-0006: Rotated dose maps simulated turned the opposite way

RadMap hands Geant4 a dose map's rotation as three angles, from which Geant4 builds a rotation matrix for the map's box. Geant4 uses that matrix as the rotation of the box's frame, which turns the box itself by the inverse. Solids reach Geant4 by another route that has always allowed for this; dose maps did not. RadMap now hands Geant4 the inverse rotation for dose maps too.

D-0007: Dose maps turned close to a quarter turn about Y simulated up to 1° off

RadMap hands Geant4 a dose map's rotation as three angles about X, Y and Z. When the second angle came within about 1.15° of ±90°, the conversion set it to exactly ±90°; on the negative side it also turned the map half a turn about its own Z axis, which maps a box onto itself. The angles are now computed exactly at every orientation.

D-0008: A full turn of a tube or sphere entered in radians left a slit

RadMap took a tube's or sphere's sweep exactly as entered. A sweep within 0.05° of a full turn, or of the pole for a sphere's theta, is now taken as exactly that, so 6.283 and 6.284 rad both give a closed solid.

D-0009: A material with its density in kg/cm³ could not be simulated

RadMap handed Geant4 each density in the unit the user chose, by the unit's name, and Geant4 knows g/cm³ and mg/cm³ but not kg/cm³. RadMap now hands Geant4 every density in g/cm³, converted. It also no longer offers kg/cm³, in which every real material is a small fraction -- osmium, the densest, is 0.0226 kg/cm³ -- and reads a library holding one as g/cm³, the value a thousand times larger.

D-0010: A solid's material could change to one whose name differs only in case

The properties panel looked up a solid's material in its list of materials ignoring case, while material names are case-sensitive, and wrote the material it found back to the solid with the solid's next edit. It now looks the material up by its exact name.

D-0011: The 3D view's arrow keys could act as if Ctrl or Shift were still held

The view kept its own record of which of Shift, Ctrl and the arrow keys were held, and cleared a key only when it saw the key released. A key released while another window had the keyboard stayed held in that record: after Ctrl+S, Ctrl+L or another shortcut that opens a dialog from the view, the arrow keys tilted the view instead of moving it, until Ctrl or Shift was pressed again. The view now reads Shift and Ctrl from each key press, and forgets the arrow keys when it loses the keyboard.

D-0012: A name in a non-Latin script prints as question marks in the PDF report

The report is set in Helvetica, one of the PDF standard fonts, which covers Western European letters only. Any other character in a project, dose map or source name, such as Greek or Cyrillic, is printed as a question mark, though RadMap accepts it as a name and shows it correctly everywhere else. The report still supports only these characters, as a deliberate limit, but RadMap now names the affected text when it creates a report.

D-0013: Full chain on one source also applies to the other sources

The setting is applied when a decay starts, and a flag recording it was set but never cleared. After the first decay of a source with Full chain on, every later decay on that thread followed its chain, whichever source it came from. Results of such projects made with 1.x or a development build should be recalculated.

D-0014: A source in a solid with "0x" in its name is not confined to the solid

Reading the geometry, Geant4 cuts every name at "0x", where its own file format puts a pointer. RadMap then asked for the volume by its full name, Geant4 did not find it, reported that only in its log, and sampled the box unconfined. Results of such projects made with 1.x or a development build should be recalculated.

D-0015: Materials named alike up to "0x" are simulated as one material

Geant4 cuts these names at "0x" where it defines them as well as where it uses them, whatever it is told about stripping names. The fix for D-0014 covered solids only. Results of such projects made with 1.x or a development build should be recalculated.

D-0016: A run stops when two elements share an isotope

The geometry file listed the second element above the isotope it shares with the first, and Geant4 reads the file from the top.

D-0017: Save Volume writes a file that loses solids when names repeat

Save Volume did not check names, as a project save does. Solids are found by name on loading, so of two with the same name one took the other's place.

D-0018: An isotope's Z could be edited to one its element cannot hold

Editing an element checked that its isotopes share one Z, but editing one of those isotopes did not.

D-0019: A source in the world volume is not confined to it

Geant4 places the world volume itself, under a name of its own, where every other solid is placed by a line of the geometry file under the name RadMap gives it. RadMap asked for the world by the usual name, Geant4 did not find it, reported that only in its log, and sampled the world unconfined. Results of such projects made with 1.x should be recalculated.

D-0020: A field marked invalid in the material editor stayed red for the next entry

Choosing another entry, or another kind of entry, filled the fields without clearing the mark an invalid value had left on them, so the new entry's valid value showed red.

D-0021: Argon and potassium had each other's atomic weight in the default library

The two entries were swapped in the library RadMap ships, from which every new project takes its own copy. Existing projects keep the library they were saved with: correct the two atomic weights in the project's material library, or reset it to the default library.