Sample polyominoes

I have recently got into writing programmes in Microsoft Visual Basic Pro - just for fun really as it twists the neurons in ways that are quite different from normal life! About 5-6 years ago I wrote a programme on an old Amstrad PCW for a friend. The basic idea is quite simple - how many different shapes can be made from a given number of squares that are only allowed to touch each other along their sides (i.e. a corner cannot be the only point of contact). Any mirror image or rotation of each shape must not be counted as they are just the same shape moved in space in some way. Figure 1 shows all of the possible combinations for 1, 2 and 3 squares after this the number of patterns increases dramatically and some of them can be quite pretty. The Amstrad is a curious beast as Mallard Basic (the free Basic interpreter provided with it) had some very powerful routines for writing indexed databases - something that's usually a bit tricky in Basic. Anyway I wrote the programme but discovered that the number of combinations gets pretty large and beyond about 11 squares I ran into severe problems of (1) speed - it's a bloody slow computer, and (2) disk space - limited to about 700kb as it doesn't have a hard drive. The initial (rather poor) solution was to get a CP/M emulator for my 486dx4-100 and run the Amstrad programme under that - this worked but was still terrifyingly slow as the emulator seemed very keen on emulating not only the instruction set but also the SPEED of a Z80 processor... Anyway to cut a long story short I have ported the programme across to Visual Basic and it now works a lot faster but there are some unexpected bugs due to the fact that the Microsoft Access database format is much less flexible than the Jetsam database used in Mallard Basic. The main problem is that it seems to have difficulty dealing with indexes composed of 8-bit characters (I use all 256 ASCII codes to encrypt and compact each shape into the database). I'm working on this and when I get it fixed I'll post the first beta version of Polyominoes for Windows here.

So I guess only two questions remain:

Well this is easy really - because it's quite a complex programme to write and the person who set me the original challenge used to sit up all night with a friend trying to draw them out on graph paper and so I said that a computer could generate the patterns rather faster and more accurately. It can - but only up to 11 squares on an Amstrad before it develops brain-lock (but the graph paper approach never got past six).

Well that's just what I was told these shapes are called!

My next project will be to see if I can figure out a mathematical relationship between the number of square in the shapes and the total number of patterns that can be generated. In theory this must be predictable but I guess it would help to know how many patterns there are first!

Edward James Oakeley