Semantic Tectonics ///
by Florian Schmidt
The Competition 2.0 – Reinventing the Virtual City issued a call to rethink the virtual city and to find new approaches for action to give positive shape to the shrinking of Second Life (SL). The invitation posits two premises that should at least be questioned.
First, it is important to note that SL is not shrinking at all, but continues to grow, though the growth curve is meanwhile no longer exponential, but linear. The astronomic number of all avatars ever created in SL, which the common media like to cite, can be ignored; but the number of concurrent users – those using the platform at the same given time – is an important figure that continues to grow. This number is currently at about 45,000 and thus twice as high as in the period of unreflected media hype in Spring 2007.
But there is no doubt that large parts of this synthetic world feel deserted. Apparently, the platform’s users do not find their way to each other in their Second Life, but are separated from each other by expansive, architectural data garbage. For in contrast to our accustomed physical world, land is not in short supply in SL. The extent of the land mass, and thus of the constructed surface, grows faster than the population. Part of the reason for this is that, in a short phase of initial euphoria, many inhabitants and companies build large buildings, but then never return. The population is subjected to an extreme fluctuation of immigration and emigration, while the number of domiciles actually only increases. So there is no shortage of residents, but a surfeit of built-up space. This situation becomes a problem primarily because Linden Lab’s world has no structure that could enable one to sensibly and intuitively orient oneself in the jungle of Potemkin facades. Up to now, we are confronted by a felt, rather than a true depopulation. But this reduces the fun in playing so much that it would hardly be surprising if an exponential exodus indeed soon followed the exponential growth. So strategies are urgently needed for this scenario.
The second premise is the starting position that Second Life is a city. Indeed, the individual sites within the simulated land surface display many things in common with our cities in the physical world. At any rate, the individual digital master builders strive mightily to borrow the metaphors they use from the field of house or settlement construction. Along with an unsurveyable number of shopping centers, Berlin’s Alexanderplatz is present in four versions. Nonetheless, it is questionable whether the resulting buildings, and in particular the world as a whole, can be understood as an urban space, because in large parts of it there is no context and no real infrastructure. But the latter is essential to our current definition of the city.
SL is much more like an archipelago state in which the individual places enter into no connection with each other at all, but drift in a maritime Nowhere, independent of each other and without visual contact with the next island.
But chaos reigns also on the mainland, where everyone is who cannot afford his own island. Where the user could expect a network of paths and a meaningful neighborliness, a crude confusion of rank urban growth reigns without center or periphery, its appearance oscillating between slum and a settlement of small garden plots, but so far hardly deserving the name city. Since the real estate in SL is completely in private ownership and every single place can also be reached by air or teleportation, none of the residents are willing to give up any of his land for public streets or paths. Neighborly cooperation can be observed in isolated cases, at best; we cannot speak of a network, and, understandably, no one is interested in a planning body. So for each individual inhabitant, it is rational and convenient to immure himself behind screening walls adorned with photo wallpaper and to block out his neighbors. For the virtual world as a whole, at least for the mainland, this understandable behavior leads to an urban-planning worst-case scenario. The runaway turn toward slums has reached dimensions threatening the existence of Linden Lab, and the solution cannot be to put SL’s residents in a straitjacket of regulations; so there is indeed an urgent need to reinvent the virtual city.
This exposé is based on the observation that, because of the fragmentation of the contents in SL, the three-dimensional web’s potentially most important advantage over the conventional Internet does not emerge at all. Essential to Web 3D, and deeply familiar to us from the physical world, is the principle of proximity and distance. Various distances between people and objects are significant and meaningful and provide us with orientation. The two-dimensional Internet is primarily oriented toward the most rapid possible, targeted acquisition of information. All points are equidistant – which means there is no space. The visitors to a virtual world, by contrast, want to find themselves embedded in an environment. They are not looking for information, but for experiences and acquaintanceships. The point is to stroll, to drift, thereby discovering things that one wasn’t explicitly looking for. But since the only effective means of transportation in SL up to now is teleportation, the spatial proximity of two places no longer plays any role. The great strength and innovation of Web 3D lies fallow. He who strolls bumps into a wall.
The concept of Semantic Tectonics therefore intervenes very fundamentally in the structure of SL by ordering space dynamically and directionally, rather than statically and coincidentally, as up to now. In contrast to a surface use plan that is firmly laid down in advance, Semantic Tectonics orders the space retroactively. The goal is to develop a self-regulating system that continuously re-orders the extremely heterogeneous and fragmented regions on the basis of their respective contents. This is not carried out by any expert team of designers, but by an algorithm that should ensure that the free play of forces in SL is enhanced, instead of stifling it with too much planning. The intervention plays out on the macro-level; the aim is to ensure the preservation of the lively and suspenseful burgeoning growth on the level of detail.
The conceptual model for the desired restructuring is taken from plate tectonics, which has shaped the form and position of the continents on our planet.
SL is well suited to this spatial concept because the virtual world already consists of loosely distributed small islands floating like ice floes on the sea; it would be no problem to integrate them into the new system. The few larger continents will be broken up into a number of smaller, independent plates, as happened long ago with the great supercontinent Pangaea. What is new is that these plates will no longer be anchored firmly in a particular spot on the map, but can drift freely over the surface of the disk world. This means that the coordinate system will lose importance, because it will no longer be possible to find sites using absolute numbers; but a search function based on content is much more powerful. The search currently implemented in SL, unfortunately, is so one-dimensional that it is useless. By contrast, the well thought-out Google page-ranking shows that technology can meanwhile provide strikingly differentiated results. The Semantic Tectonics will be based on an algorithm that resembles a complex search engine. This formula is the motor, the driving power that regulates the direction and speed of the continental drift. The most important influences are the ranking, i.e., the qualitative evaluation of individual regions, and the so-called tags, keywords that enable a semantic categorization. Following the principle of a folk taxonomy (folksonomy), each resident has the possibility to assign values and terms to individual sites; these values and terms are input in calculating the process of transformation. In this way, the topical emphases of the respective plots of land can be ascertained. A great number of possibly contradictory descriptions can thereby be assigned to one plot of land. For example, the fictional Shrinking Cities islands could be tagged with the terms architecture, urbanism, culture, conference, discussion, book, art, design, etc. and then, depending on the frequency of the tags, take a corresponding course. After the Semantic Tectonics take effect, all land masses in SL and the words will successively order themselves in a tag cloud around certain core themes. The new world map contains a number of widely separated “magnetic poles” that are the respective epicenters of superordinated themes. There will surely be an eroticism pole, toward which will eventually move all islands offering explicitly pornographic content. Not all too far away could be the gambling pole, around which a floating Las Vegas gradually arrays itself. Between these two centers, a field of tension will arise in which many sites of today’s SL will find a new home. Research and instruction, for example, form a continent on the other side of the world, surrounded by a chain of islands – the individual colleges and universities. Extreme subcultures like the furries could contract their enclaves, until now widely scattered, into an “empire of the fur-bearers”, where they could indulge in their role-play completely undisturbed.
Important is that the regions are not taxonomized from the top down, but that the inhabitants of each piece of land make their own decisions on a small scale, which then have effects on the greater scale, developing their ordering power. The system is grass-roots democratic and scalable without problem.
Of course, this new spatial strategy will also lead to conflicts. As with the real continental drift on earth, trenches and cracks will form, presumably along the predetermined breaking points of the existing SIMs. Not every inhabitant will consent to see his island suddenly moving in the direction of the eroticism pole, just because his neighbor has a predilection for lacquer and leather. Vice versa, it is conceivable that precisely this neighbor is the first to give the settlement the right spice, and that one doesn’t want to lose him at any price. For this reason, each owner of a plot of land should have the possibility to establish firm bonds with his neighbors, so that, if all sides agree, plots of land cannot be separated from each other. In addition, everyone is free to move to a region that is more moderate, and thus also more heterogeneous. But completely uniform neighborhoods would arise only around the centers of thematic poles, i.e., among people who dedicate themselves completely to one thing and who thus are glad to be among their own kind.
Since we are dealing with a world whose surface is covered mostly by water, integrating houseboats and other water-going vehicles could meaningfully expand the concept. In contrast to islands, these could be voluntarily steered by their residents and combined with each other in temporary settlements. The breaking up and drifting apart of the main continents would give rise to new waterways, through which the boat owners could navigate. As soon as Semantic Tectonics developed its full effect, there would therefore be essentially three different settlement regions: floating cities consisting of boats; drifting cities cut through by canals, which would resemble Venice in some ways; and the great agglomerative centers on the continuous masses of solid land.
For the latter, it would be extremely useful for Linden Lab to lay out a few main traffic arteries that would divide the space into comprehensible areas and make strolling possible. Public squares for the respective communities would markedly enhance the quality of life and counteract emigration and urban decay. The resulting public traffic network would form a kind of central nervous system for the virtual world and could be easily financed by the operator, since the plots of land close to the traffic arteries could be sold for much higher prices than a bit of nothing out in nowhere. The SL server architecture also requires a fundamental overhaul.
In the future, frequently visited regions should be assigned a computer capacity proportional to the number of people visiting them. Such a flexible load balancing is urgently needed to end the phenomenon of popular sites having to be closed because they are too full as soon as several dozen avatars arrive, while elsewhere computing capacity is wasted to calculate fallow areas. Since the principle of Semantic Tectonics evaluates the individual regions anyway, the same key could be used to elastically adjust the technological support. Deserted or neglected plots of land drift toward the periphery, where ultimately they will no longer unnecessarily burden the patience of seekers or the capacity of the servers.
Virtual worlds offer a wonderful playground, not only for mental experiments, but also for urban-planning interventions with very real social feedback loops – for if the virtual city is designed unattractively or structured poorly, the inhabitants simply emigrate. Without a doubt, these accelerated processes can teach us a lot that can be helpful in the real world, as well. Vice versa, when intervening in synthetic worlds, we should make use of all the freedoms that are given to us only there and should not unnecessarily limit ourselves by reproducing the restrictions of the physical world.