Current Version 2.4.2
Some fixes to several annoying bugs that have been found and a few usability features have been added. The most significant change is to the handling of adjustable step lengths in Runge-Kutta solutions.
- Some logic keywords were not colouring correctly in the edit view;
- The power operator (^) was occasionally causing an incorrect syntax error;
- Improved error handling in the calculation of chi-sq % in the merit procedure;
- Improved error handling for parameter values close to zero in the Marquardt procedure;
- Graphing bugs which led to floating point errors and mishaps when re-loading graphs have been fixed;
- An interference between output and integration steplength for adjustable Runge-Kutta has been resolved. This affected some models where integration step length changed significantly over the course of the simulation;
- An occasional bug when arrays currently shown on graphs and tables are re-sized has been fixed;
- Improved handling of missing grid files.
- The symbol trace form has been modified so that if a symbol is selected in the edit view it is automatically selected in the trace form when opened;
- The facility to convert a symbol (e.g. from variable to ode) has been added.
Hopefully there is no reason why you would want to download and use an old version! But just in case they are available here…
Version 2.4.1 was a A modest update with improved example models and examples guide. There are also some fixes to small but annoying bugs which make it worth upgrading from 2.4.0.
- fix for occasional errors saving graph and table data;
- uncertainty analysis extended to allow use of arrays;
- an error in calculation of the Nash Index for small RSS values fixed;
- number formats improved for small values in some dialogs;
- more consistent labeling for the replacement variable dialogs;
- bugs on the parameter sample dialog fixed.
- improved proof reading of the user guide!
Version 2.4.0 has benefited from an overhaul of the software and documentation which now includes enhanced examples and step by step illustrations on how to apply OpenModel.
Thanks to some (modestly) funded OpenModel work we have been able to obtain a little bit of software development help to oversee improvements to the underlying code. As a result I believe it is more stable and consistent to use, but fear not, it will probably still crash and burn occasionally (hopefully not often). For better or worse it is written by scientists, not software developers…