Imagine life in a warm climate, enjoying what one might call the blessings of the sky. Such open-to-sky spaces are of crucial importance to architecture — not only for the subtle and metaphysical feelings they engender within us, but also because of the decisive role they can play in the creation of humane (and affordable!) habitats. Today, in many Third World cities, the majority simply cannot afford the kind of housing that is being produced. It is time to go back to fundamentals. We must learn again the invaluable role of open-to-sky spaces in enhancing habitat, while drastically reducing their cost. For instance, a courtyard is used for a number of essential functions e.g. cooking, entertaining friends, children’s play, and so forth. This has always been an integral part of the indigenous housing typologies produced in from the Kasbahs of North Africa to the hill towns of Italy. How do we integrate such spaces into the new housing fabric? Here are a few cues to develop a reasonably high density of 500 persons per hectare (including open spaces and social amenities), and yet give each family an individual house on an independent plot. Instead of stringing out the houses along a street, these can be grouped in clusters around courtyards. The basic module may consist of a number of houses arranged around a court measuring less than a katha each. We can form a larger cluster of up to two dozen houses arranged around a larger open space by repeating thrice the basic module. The larger unit in turn can be repeated thrice to form an even larger open space- and so on, right up to the central open area- the focus of the whole community. The houses are only starting points; each of them can be extended and embellished as the family wishes. Architecturally, they are but a simple kit of parts. The plots on which the houses stand are nearly all the same size, varying from 0.6 katha (as in DUIIP project in Mirpur) to 1 katha. The scheme can be designed to cater for a wide income range, the plot size variation should be kept minimal to ensure that equity is built into the layout. For a poor family who might be able to afford only a lean-to, the rest of the plot provides the open-to-sky space rendering the habitat liveable, e.g., provides a place for a shady tree or a cow. The better-off families can use the same plot to build town-houses. Most London terrace houses are only about 6 metres wide; Amsterdam canal houses are often much narrower 3-4 metres, as are houses in traditional old Dhaka settlements. In order to make the design truly incremental, each house should stand free of its neighbours. This ensures the essential cross-ventilation, and allows it to grow independently of the surrounding houses. At the same time, in order to economise on land, each house touches the plot boundaries on two adjacent sides. Windows are allowed on the other two sides, and also on the side, which abuts the central courtyard. In short, this pattern generates a hierarchy of open-to-sky spaces, from the most private realm to the most public, all within the context of house typologies, which are incremental- and malleable. That is a habitat in which people can add their own layers of meaning, giving personal and cultural identity to their habitat. This is the traditional way in which people have humanised the houses they live in, all down the centuries. This is precisely what is impossible for them to achieve within the context of contemporary housing, because the buildings seem so intimidating, so immalleable. Malleability is indeed a quality crucial to successful habitat. I would now like to look at shelter from the whole city context. I can feel the frustration of many stemming from the fact that they are forced to deal with the symptoms of a much graver problem — the appalling scarcity of urban land. For instance, Dhaka, a city of 10 million, has the same elongation tendency towards the north it had 200 years ago with a hundredth of the population. That is, while demand has multiplied several-fold, supply is still a trickle. The result: galloping real estate prices and ever-increasing numbers of basteebashis (no less than 3 million). So, it is essential that we find ways, in each of our cities, of increasing the supply of urban land on a scale commensurate with the demand. Urban land means access to jobs and public transport, hence re-structuring our cities. This is what Dhaka is about. Opening up Yusufganj or other satellite towns could be attempted to modify the North-South linear pattern into an East-West structure. These could be self-sufficient new neighbourhoods of up to half million people. The structure plan shows how to use public transport to open up land. Now, on the scale of half million people, one has the opportunity to perceive the fundamentals. For instance, what kind of housing should we build given the income profile? Or, granted that courtyards and terraces provide additional room, can we really afford this amount of open space? In other words, don’t high-rise buildings save a lot of land for the city? Actually, only about a third of the land in a city is devoted to housing including neighbourhood roads. So, piling up people on a particular site does not save much land for the city. But, it does deprive them of the crucial benefits of the open-to-sky spaces. For decades now, architects and engineers have been searching for miracle new materials and technologies to produce low-income housing, when, all along, the land-use planners have stated the question wrongly to begin with. The problem of low-cost housing is not like the medieval alchemist’s fevered hunt for the force to elusive touchstone that would turn dross into gold. On the contrary, it is primarily a matter of opening up the supply of urban land, of identifying optimal densities, and of re-establishing land-use allocations between the diverse functions that make up a city. This brings another crucial issue: typologies. Dhaka could almost be any one of a hundred cities in Asia, Africa, or South America. The rich live in high-rise buildings, the poor in shacks which travel like a river of poverty through the spaces between these towers. What is frightening is not just the contrast of income levels, but the vicious discontinuity between the concrete and steel towers where the rich live and the polythene lean-tos of the squatters. We have lived with rich and poor — as has much of the rest of the world. But, the difference was that, regardless of the differences in sizes and materials between palaces and the most humble abode, there was certain continuity in their typologies. In contrast, the grotesque discontinuities we observe in our urban centres today are a horrifying portent of the breakdown within the social fabric of society itself. I will end with two images. One fills me with great hope, the other with a kind of despair. This man lives in a miserable existence in an unused sewer pipe — sharing a cup of tea with a friend. It’s a social occasion! Humanity in the Third World is still intact. In the ultimate analysis, that’s probably the greatest strength of all. The other image is a sadder one of the high-rise towers in Eskaton or Banani. Silhouetted in the foreground are the squatters and construction workers. Behind them rises a bunch of new skyscrapers. Many of these buildings are deplorable. But to them, it is the surreal, mythic image of the city, which they yearn for, but can never attain. The discontinuity of built-form is truly horrendous. Until we find ways to change this, there can be little hope. Mahbub Rahman PhD offers settlement and design courses at the BRAC University and the University of Asia Pacific, and advises the State University of Bangladesh School of Architecture & Design.