CHAOS, DATABASES and FRACTAL DIMENSION of REGIONAL ARCHITECTURE Jadwiga C. Zarnowiecka PhD in Architecture Faculty of Architecture TU of Bialystok Krakowska 9, 15-875 Bialystok tel/fax: 85 422929, tel. 85 423214 w.127 e-mail: zarnow@cksr.ac.bialystok.pl The greatest defenders of new science state that the 20-th century will be called a century of relativity, quantum mechanics and chaos. (Michael F. Schlesinger) Modern research on chaos started in the 60’s from an incredible finding that simple mathematical equations can model systems as complicated as waterfalls. In the 70’s some scientists in the USA and in Europe started to find their way through the chaos. They were dealing with different spheres of science: mathematics, physics, biology, chemistry, physiology, ecology, economy. In the next 10 years’ time the term ‘chaos’ has become generally known in science. Scientists gather in research groups according to their interests as to chaos and secondly according to their scientific specialities. (Gleick 1996) Objects that described chaos were irregular in shape, ripped. In 1975 Benoit Mandelbrot called them fractals. Fractal dimension that described fractal objects was also his invention. Fractal dimension is a way to measure quality: the degree of harshness, uneveness, irregularity of a given object. Carl Bovill (1996) showed how one can use fractal geometry in architecture and designing. This very fact made me try to use fractal geometry to deal with regional architecture. What or who is the degree of regionality of a given object to be for? A specially qualified person is able to state it nearly automatically. However, regionality is in some sense an unmeasurable feature. While dealing with data basis or checking particular projects, creation of procedures of automatic acquiring information concerning regionality is becoming a necessity. CHAOS Even though it was already in the XIX century that Henri Poincaré noticed the outlines of chaos, the first confirmation of its existence was done by Edward Lorenz in 1963. Lorenz was carrying computer research on weather. He wanted to clarify why there were discrepancies between the weather forecasted and the real one. He made a mathematical model of atmosphere composed of 12 equations. The system of equations was beeing solved by computer. Accidentaly, to speed up the calculations, he introduced the intermediate values, which he got in a previous simulation. To make it easier, he rounded off the values from 6 to 3 places after decimal point (see Table 1). During the simulation of two months of weather this initially small difference of results became as big as the very signal from the beginning. In such a case if the real atmosphere behaves in the same way, it is impossible to forecast the weather months ahead. Even the most powerful computers won’t be of help here; small mistakes would be growing to get big. The effect of sensitivity to the initial conditions was called ‘the butterfly effect’. (It comes from Lorenz’s publication ‘Can the flap of a butterfly stir up a tornado in Texas?’ In a TV interview Lorenz explains that the title concerning the sensitivity to the initial conditions was chosen by the publisher, and