An experiment: take an elastic band and hook it over your thumbs, then stretch the elastic by moving your hands gradually apart. At first the band will move easily, but the tension will gradually increase — keep watching closely. Eventually the elastic will start to become lighter in colour and small bits of the edges will tear away, leaving semi-circular indentations in the edge of the elastic. Further gentle pressure will result in one of these local tears suddenly ripping all the way across the band and it will then break — very quickly.

This behaviour is typical of many real-world events. Things proceed along well-defined patterns of behaviour (the elastic band stretches as you pull harder) but then suddenly change what they are doing and rapidly do something completely different (the elastic snaps!). Whilst it is easy to get measurements for the well-behaved situation, it is much harder to predict or work out what is happening for the unusual sudden change. Because of this science and mathematics have had rules governing how many things behave, but only recently has an approach been developed which can cope with describing what happens when something is breaking down and rapidly changing. The mathematics is called Chaos Theory. In fact, the most popular image of Chaos, the Mandelbrot Set, is just a diagram showing how an equation can suddenly behave peculiarly.

In case you were wondering if this really is the Atari Notes page (note the change: ST Notes seems less appropriate now that the Falcon is upon us), imagine a world where the development of computers was chaotic (I said imagine!). So in this world you could imagine a computer like the Atari ST, which gradually develops as the 512k base model becomes the 1MB RAM version because the latest software needs it, where the palette of colour increases from 16 at once on the screen to 4096, and where solder-in RAM chips are replaced by the readily available (and cheap) SIMM modules. The software would echo this gradual hardware development, so the initial releases would just be conversions of earlier 8-bit programs, but would soon begin to exploit the capabilities of the 16-bit processing power of the ST, producing first vertical scrolling, then horizontal scrolling, and finally multiplane horizontal scrolling, and ultimately some very sophisticated modelling and vector techniques.

Music software in this hypothetical world would be carried along with these computer developments too. Early simple tape recorder emulations would evolve into complex and elaborate systems, fine-tuned by user feedback into very efficient music-producing tools requiring at least 1MB of RAM, and probably considerably more, as well as at least a hi-res mono screen if not a large screen monitor.

Graphical and user interface sophistication would increase, resulting in powerful grid editors and score/notation based editors, and even object-oriented approaches to the viewing of the musical data. Hardware add-ons would enhance the capabilities of the computer and make it more suitable for serious musical purposes: more input/output ports, synchronisation facilities etc.

But then suddenly, this cosy world would begin to fall apart. Tell-tale little signs like games manufacturers announcing that there would be no more releases on the computer perhaps, or even a sudden move by software writers to port their software to other types of computer platform. Programs that would once have appeared first on our hypothetical computer start to appear first on other machines, with promises for it to be available soon on 'other computers'. The neat and cosy, predictable world of the stretching elastic band is about to be replaced by the chaotic breakdown as it starts to break apart; except that here we are talking about a computer that is ending the end of its product lifetime.

At this point, an interesting thing often happens. The computer manufacturer releases a radically new and innovative product that makes the existing one look like it belongs with flared trousers and tank-tops. Suddenly no-one wants to know about the 'old' product and lunges to get the new one — and the elastic breaks. It isn't a pretty sight, since the two bits of the band wiggle around for a while until they eventually get dropped, but whilst they are wiggling you find that the magazines and some software producers continue to support the computer, insisting that it is still OK. Sharp price drops encourage the not-so-well-informed into buying what looks like a bargain, only later to discover that they should have waited.

The new 'all-singing, all dancing' machine quickly becomes the one to buy, and everyone updates their software to make it compatible. Magazines stop mentioning the 'previous' model and start to rave about the new model. Soon, any thoughts of even testing software on the 'abandoned' predecessor are abandoned and the new machine begins the same process of evolution and sudden decline all over again.

Back in the real world, you might be able to see some similarities to what is happening to the Atari ST now that the Falcon030 has been announced. Similar things have happened to others: the BBC Model B disappeared when the Master series was announced, and then the Archimedes killed almost all the previous Acorn computers. Apple products can suddenly change their complexion — when the Mac II was announced, the classic one-box Mac looked small and restrictive almost overnight. But this characteristic behaviour is not restricted to computers — look at what happens to hi-tech musical instruments like synthesizers or drum machines. A new release gets a rave review, and yet a year or so later is sold off cheap as the latest versions are announced. Many types of recorded music have a fixation with charts, where old material is cynically relegated to the scrap heap on a weekly basis. Perhaps computers are lucky that they seem to last for up to five years or so before they are 'out of fashion' and 'superseded'.

You now know the danger signs to watch out for. You may be surprised at how effective even a little knowledge like this can be. But applying it to the real world is hard. After all, human beings are chaotic too.

As a parting thought, remember that styles of music undergo exactly the same process of initiation, gradual development and rise in popularity, then the catastrophic plunge into obscurity.

AND NOW...

Operating Systems also evolve. TOS 2 has several features that the 1.n versions lacked: it offers keyboard shortcuts for the desktop functions at last, and lets you have lots of icons and lots of colours as well. MultiTOS promises to allow pre-emptive multitasking for desktop publishing and other 'more than one program at once' power users (I have not yet seen any music programs working under MultiTOS, so I reserve my judgement for the moment). But all the STs so far have been let down with a somewhat pedestrian way of handling fonts. The basic ST has the system font, and anything more sophisticated needs to be added on by software.

A standardised way to handle multiple bitmapped fonts does exist: GDOS, an extension to the ST's GEM operating system which provides multiple bit-mapped fonts and handles all sorts of graphics output devices (including the screen). GDOS has a reputation for awkwardness and slowness - and it has been paused at V1.1 for several years, apparently waiting for FSMGDOS (based around scalable outline fonts, but still awaiting release). FSMGDOS should finally allow the mixing of bit-mapped and outline fonts, which should please everyone, help remove some of the confusion that abounds in connection with fonts, and when used on the Falcon, perhaps even give Macintosh users something to think about!

LARGE SCREEN?

In all the euphoria which will no doubt surround the Atari Falcon030 over the next few months, one interesting observation may well get overlooked. Remember that despite the powerful Digital Signal Processing, fast processing speed (about four times that of your basic ST), cheap PC VGA monitor compatibility and up to 65,536 colours on the screen at once, there is one drawback. The ST high-res mode provides a 640x400 pixel black and white display, just like the existing ST, but the only mode with any more pixels is the Super VGA mode, giving 768x480 in 32,000 colours (this is one of several SVGA screen resolution/number of colours combinations). Those extra pixels are obtained at quite a heavy price, however, and the higher number of colours means that more memory needs to be moved around than in the 640x400 monochrome mode — so you can expect a drop in screen update performance just to get a slightly larger screen.

I would suggest that colours aren't much use for a computer which is intended to produce music using MIDI, and hundreds or thousands of colours is definitely overkill. At most eight or 16 colours would be enough to enable different note values, controllers etc. to be easily seen. What is needed for serious musical applications is a large screen, especially as sequencers get more sophisticated. Adding a larger screen with an add-on board is likely to increase the price of buying a Falcon, especially since you will then probably need a new large screen monitor, or perhaps support for another of the SVGA screen resolutions (1024x768 is possible with SVGA on the PC). This is not new — apparently the expansion busses on the Mega ST, TT and Mega STE have mainly been utilised by graphics boards.