Obviously this project won't be around for ever - things built from Lego or fischertechnik generally aren't - they are eventually dismantled and the parts recycled, and that is one reason why I now like to record them in web pages while I have them. Also, it takes up a fair bit of shelf space.
I am planning a Version 2 of the control program, however. Each draught goes through a sequence - identify, push onto belt, wait, push off belt - and the program deals with one draught at a time, which makes the programming very straightforward. Several people have suggested that it could "multi-task", ie deal with more than one draught at a time.
The minimum time between successive draughts would be the time taken to travel between the two pushers, and this would allow up to three draughts on the belt at a time, each of which could be a black draught, a white draught or no draught. The program will therefore have to deal with three overlapping sequences, which could all be different, at any given time, while also maintaining the belt counter and knowing how far each individual draught has travelled.
I think I can work how to do that, and it will be a more interesting program to write. It will need interrupts, to manage the belt counting, and the interrupt facility in PICAXE Basic is very limited, but I think it is doable.
Something else that I would like to try is a more intelligent way of using the light sensor, taking readings with and without the infra-red illumination and calculating the difference in order to minimise the effect of ambient light.
If it works, and I can't see why it shouldn't, this could extend the usefulness of these simple but versatile sensors when used for line following and object detection.
I don't have any plans to redesign the hardware - by and large, it works as expected. I am not 100% happy with the conveyor mechanism, however, and the belt counter is noisy, jerky and irritating, so I might see whether either of those can be improved.