Since Report 4 there has last been some real progress. I now have a BBC Computer running MicroCODIL using the BBC DFS system and old 5 inch floppy discs. I also have a modified BBC Master with a 500 MByte flash drive added and I rescuing all my old floppy discs (at least those that are still readable after 25 years) by copying the programs and data to the flash drive. MicroCODIL will need some hopefully minor modification to run on this system as the software is designed to run from drive 0 - which is the drive number for the flash drive. However it should soon be possible for me to supply copies of the software to people who have access to a BBC or BBC Master Computer.
Work on transferring the software to a P.C. - using a BBC Basic interpreter - has not yet got underway. I will need to see how practical it is. The problem is that only a small part of the program does the "real work" and much of the code relates to providing a good user interface in the context of the BBC computer (allowing for various models) using a 40 column by 25 lines teletext screen layout, which the interpreter does not allow to be changed. I am therefore looking at the possibility of producing an experimental subset modified to simulate running on a neural network model - which will be experimentally very exciting. However such a rewrite might need to be done by someone more familiar with modern compilers and interpreters..