This is a very interesting style. Not because the paintings are better or worse than any others, but because it relates to all sorts of things which can call into question lots of assumptions we make. Let's look at three computerish paintings by Albers, Noland and Mondrian.

On the left, Homage to the square by Albers strikes us as painted according to a recipe. The colours become lighter and increasingly saturated as they go in to the centre, with the hue varying in a less obvious manner - perhaps from redder to yellower - perhaps not very much at all. What about the sizes and positioning of the squares? The squares evidently reduce in linear dimension by exactly the same amount - (40 pixels in my version). The positions are arranged so that each square drops by thirty pixels, leaving ten pixels spacing at the bottom - which adds, as it must, to make the 40 pixels. Anyone with a smattering of knowledge of colour and trivial mathematical skill could write a program to generate things very like Homage to the Square in a half-hour or so. Some questions arise from this, though. Firstly would Albers have done this computerish painting (colour field painting, in art style terminology) if computers had been around and as easily accessible when he did this in 1964? Does the computer do for this sort of painting what the camera did for normal painting? Even if we assume that the paint has been laid on as evenly as possible, and that the measurements really do follow the simple progression suggested, there is still plenty of room for the artists choice. There is the size of the first square; the decrement to its dimensions in producing the others; the number of squares to be included; the amount by which the brightness of the squares changes. There is also the amount by which the saturation changes (if at all), the amount by which the hue changes (if at all) and the amount of the centre offset that gives the asymmetry. Even if we ignore the saturation and hue variations, and give a very conservative estimate of say 20 for the number of possibilities for each of the other choices, we have a possible number of paintings like this of about three million. How many of them would be as effective as Homage to the square? Apart from this, Albers had to choose to use squares, to use the changing position of the centre, to vary the hue/saturation/value parameters. Yes, he made a lot of choices.

Look now at Gift, by Noland (right). Superficially it could almost be seen as a circular version of Square. The basic colour is similar. The same simple set of geometric figures placed one upon the other is seen. In fact though, it has obvious differences. The discs are concentric. The circles contrast with the square background. And there is no sign of the mathematical rules. The artist has not operated within such an obvious set of self-imposed rules. Does this make the painting better or worse - or is it immaterial? Is the painting more computerish, with its concentric symmetry, of less so in view of the absence of mathematical progressions of size and colour?

Our last computerish painter - and maybe the most famous - is Mondrian. His works seemed to influence a generation of designers, and hence the new buildings that were put up in the fifties around London to replace those lost to bombing. Many had the bright red, yellow and blue squares of colour placed around them in a calm but colourful attempt to brighten up schools, swimming pools and colleges. Some still exist, and the power of the Mondrian-like colour layouts to transport one back to the fifties is immense.

One could write a program to generate Mondrian-like paintings, but it would be far less effective than a Noland painting program, or an Albers painting program. And in fact if one traces Mondrian's work as he developed, one can see that it changed slowly, and that at no point does he seem to have suddenly "gone abstract". The post-painterly abstractionists, like Albers and Noland produce work that doesn't seem to be referential to anything in the real world. Their work seems like a cerebral creation - left-hemisphere stuff, in pseudo-psychological parlance. Mondrian, though, seems to have transformed the real world, via a time-varying human filter, gradually producing works like Composition 10 after many years of natural development. Still - I'd like to do a computer Mondrian - just to see - although even in this period Mondrian's work was far more complex than that of Albers or Noland.

One thing that should be borne in mind, when talking of computerish painting, is that the hard-edged machine-like qualities that most normal people associate with a computer are a red herring. Anyone with a penchant for making mechanical devices could create a computer-driven brush-manipulating engine, which could splodge away in very human style, or at least like a chimpanzee!