When the NAS­COM users’ newsletter announced its competition for an entertaining program to run on a 1Kbyte machine it seemed immediately apparent to two certain people that an entry developed on their machine should most certainly be submitted. Sadly, though, they were revising at the time for finals (congratulations, incidentally, to them both) and had not the time required to squeeze any of their amazing programs into the requisite tiny space. Not for nothing, though, had they spent so much time in the company of Arthur Norman (see previous articles in DATABUS for samples of his work!) – with a pair of malicious grins almost worthy even of him, they remarked “Arthur and Chris can write some code for it”.

This suggestion did not fall on entirely stony ground. Both Arthur and I have a fairly long record as hackers of the PDP-7 and he has additional vast experience of entertaining software on many assorted machines. We both felt that the time had arrived to get fairly close to the machine code of a modern microprocessor, and what better machine could there be than the Z80? Furthermore, what could be a more amenable introduction than the challenge of writing a fun program to meet a deadline?

I must admit that we were both quite strongly taken aback to discover that the aforementioned deadline was less than two weeks away, but with mutual cries of encouragement we nerved ourselves for the fray, carrying the poor innocent NAS­COM off to to the fifth floor of the Computer Laboratory Tower and denoting ourselves the FIFth FLOor SOftware company.

One of the factors influencing our choice of program to write was the shortage of available time, not only to implement it, but even just to decide what it should be. Some of the considerations influencing our choice of a moonlander were

1. We each had some experience of looking at programs for this, respectively on the PDP-7 and the GT40.
2. It seemed to us that the problem could be expanded or shrunk more or less at will, and according to how well we were getting on with it.
3. The majority of moonlanders are BAD and BORING and we wanted to write a half way competent one.

Having determined this fundamental point, we turned our attention to the machine itself. A number of differences in philosophy from the machines with which we were familiar brought themselves forcibly to our attention. Probably the most significant of these was the bewildering variety of available instructions: the PDP-7, for example, has only a four bit opcode in its eighteen bit word, and there are no prizes for guessing how many instructions that gives it. By comparison with this, eight bits of Z80 opcodes, most of the combinations of which are legal, appeared to mean a great many more possibilities to consider.

Offset against this, we had occurences of the bewidering variety of the