Fixed bug when converting from Cross Section to Diameter. When the Mie curve crossed a boundary multiple times only the upward crossings were identified.
When a different set of cal data is loaded the GUI now clears any old calculated diameters.
Fixed bug which meant that when two bin boundaries had their three sigma error bars crossing the integration could find zero probability solutions bins with width 0 (introducing infinities into the calculations) or negative widths
Fixed bug which caused -1#ind results if two consecutive points in a Mie table were identical and a bin boundary fel on these points
Fixed bug which caused -1#ind results if when integrating over a gaussian the upper and lower guesses were not identical but gave identical diameters due to rounding errors.
Fixed bug where a cross section guess exactly at a max or min would cause problems.
Fixed bug when integration over a gaussian was not over the correct range.
Changed method for calculating diameter mids and widths based on a straight line fit to integrate over guesses in gradient and intercept
added extra columns in the displayed table for gradient, intercept and their variances and covariance
Fixed bug where cross section width error was being incorrectly exported as the diameter width error
Added output of logarithmic diameter centre and width
Added ability to add extra rows to the table
Fixed problem whereby a cell in the table didn't update if it was still being edited when the user generated boundaries
Added some extra flexibility when reading calibration files so the formatting doesn't need to be quite so strict
Updated the manual, with description of how to create cal input files and replaced theory section with reference to the paper
Added two example input files
Ensured all derived columns are cleared when new calibration data is loaded
Added the ability to use multiple columns in a scattering properties file to represent uncertainty with options to weight these different columns equally, normally or in a custom manner.
Fixed a bug where diameters were incorrectly calculated if the variances and covariance of a straight line fit gave a corellation coefficient less than or equal to -1 or greater than or equal to 1.
Fixed a rounding error which could cause the wrong gradient, intercept, etc to be assigned to a bin if a calibration file included floating point threshold values.
Added Makefile and shell script to build under linux.
Minor code changes to allow compilation with gcc.
Fixed some text related bugs to (related to Linux vs Windows newlines and string copying) to ensure correct operation on Linux.
Calculating Diameters now stops if it finds bad input that cannot be converted to numbers, indicating to the users which cells are bad. Peviously bad input was substituted for -999.0 in calculations.
Fixed bug introduced in version 1.2.8 which stopped CStoDConverter processing valid files
Improved catching of errors in CStoDConverter input
Moved some functions from cstodconverterframe files to cstod files
Created CStoDConverter console application