Digital architecture Possible definition: A species of architecture of its own. Most time sleeping in the depth of a data file on a hard disk, sometimes it has a chance to awaken to impressing, sometimes spectacular existence. It has some things in common with virtual architecture. But it is not identical with it, though. Anything not real, but happening onlay in your or my mind and imagination may be virtual. Digital architecture is not hardware, not a piece of touchable concrete. It is, taking into count the possibilities of today's computer technology, still a voyage of the mind. It can't use every input channel into the viewer's mind, because it is still difficult today to reproduce smells and tactile/haptic effects. So today's digital architecture has to use reduction of sensoral input. This hasn't to be a disadvantage. If the quality is given, combined with a minimum quality of the way of displaying it and a good, adequate presentation, digital architecture creates immersion. The user will only think about what fascinates him when viewing the design, forgetting that he or she wears weird eyeglasses, looking at a medium quality picture in a cave (digital one, not stone age). There may be differences to real architecture, e.g. there usually is nothing like digital rain, so the building does not need to have a roof. But there are things it shares with real architecture. Every building has to tell a story, no matter if it is a real or a digital one. Otherwise it will become boring. A possible "extension" of "digital architecture" to"real architecture": * After thinking a long time about what I read in the past few years about new inventions in the field of AI (Artificial Intelligence), a kind of intelligent agent which originated in a computer game fascinates me more and more. Can you remember the computer game "Creatures"? There you had to breed a furry species of animal, keep them away from dangers and bring up, educate and train them. They had an intelligence of their own and even a little bit of genetics based on a model of success and failure. The AI model behind this game was so good that the British armed forces took it to develop artificially intelligent new simulators and weapons. The artificial being was put into an artificial plane mid-air. The time it was flying without crashing into ground was measured. The two or three most successful of a dozen were kept, the other ones removed from the system. The digital "genes" of those few best ones were taken as a base for the next generation, which varied them a little bit, and then was put into the flight simulator, too. With this model, there is a possibility slightly larger than pure stochastics to get from "Quax the crash pilot" (old, famous German movie with Heinz Ruehmann) to flying loopings and other complex manoeuvres in an acceptable number and time of a few thousand computer generations. Now, what does this model share with architectural appliances? A building, no matter if it is a digital one or a real one, has to fulfill certain expectations. If those can be parametrized (defined in a good/bad catalogue of parameters), the building may have a chance to evolve by using this AI model into one fulfilling those expectations the best way possible. This may be difficult for some aspects of art - because this is very subjective, but is used very successful for developing those parts of buildings carrying heavy loads - this may be parametrized rather objective. The new thing about this model is that it develops intelligently. E.g. today's applications optimizing load-bearing parts mostly use a simple +/- model: Add thickness till stable enough. Add more thickness. If there is no more effect begin to subtract again. The AI model could possibly make this process faster and more evolutionary. For reaching this aim it uses parametrization and agent-based models. This may be combined with the follwoing aspects: * Can be parametrized Have you ever seen an existing wall of a house with three switches for heigth, width and depth, which changes its shape if you touch these switches? * Changes viewable at once As soon as you have made all your changes to a design, you may take a look at them. No week-long waiting for them to be built. * Documentable heritage A data file may contain a "document history" which shows how the design developed to its current form. This is possible in classic architecture, too. But there you will find tons of sheets of scrap paper on your table - not a single file. In digital architecture, you can view it click by click. Comparable things I can think about are the z/u(ndo)/go-back button in AutoCAD by AutoDesk or the history trees of some well-ripened content and version managing systems like CVS for Linux (I know about the Pharmacy and Cervisia frontends). This is with certain similarity to the evolution of a design on scratch paper. It allows you to navigate to a defined stage of the design, examine it in comparison to others and judge and/or refine it, possibly with parts of a later stage. And this much easier than having to copy it all from at last three sheets of paper Current disadvantages: * I am currently participating a thread in the de.sci.architektur newsgroup about the current way of the first steps of designing: Do everything on screen by computer or still use a 6B pencil for the first steps? The opinion of most of the participants was that, beginning with the late phase of form-finding, almost everything might be done using a computer. But for the first phase, a pencil is the better thing. One advantage is the softness of the drawing. The other one is the wide range of interpretations possibly by this softness, or better the missing exactness. I am still waiting for this inexact computer tool... ('teddy' (-> www.archmatic.de/www.glossar.de) might be a first experiment - more still to come).