what is needed is an investigation of the potential performance and memory consumption issues. The result of this investigation should be delivered as a report, meaning written document discussing what changes have occured between the FreeMind versions. IMO writing Perl scripts is completely unnecessary for the task, as creating a mind map that has 10 MB manually is easy and quick.
I have created a perl script for test map generation. I have also
performed some memory consumption and cpu time analysis and shall report
my results later. The script for creation of the test maps could help
you too, so I post it now. The properties of the generated map can be
set in the configuration part of the script.