Publication 3


AUTHOR: Henri Comrie

A presentation of ideas and principles for the design of affordable housing, in the context of much-needed debate.

There must be many unbuilt housing projects that populate the bottom drawers of architects and urban designers across the country. They may be considered failures at one level, but valuable contributions to a reflective debate at another. We have selected three unbuilt projects to highlight our approach.

The first design was our submission1 for an ideas competition (2003 SIDA International Housing Competition), the second a public sector commission (2005 N2 Gateway Delft Symphony Precinct) 2, and the third a private sector commission (2008 Fairvalley Eco Estate for Kagiso Urban Management, in association with Mike Schroeder Architect). The last has been creatively managed and was the only project that used a design-led charette3 process.

All three designs are for green field sites and are conceived at precinct scale. The projects symbolise a trajectory of involvement from isolated visioning (ideas competition) to interdisciplinary engagement by a public sector client and by a private sector one.

In every case we very much enjoyed being involved as crafters of the fine-grained, three-dimensional urban spaces associated with affordable housing. In the latter two projects we appreciated working in balanced and creative teams and actively engaging in design-led debate.


The principles that we employed have remained relatively consistent. Through highlighting the variety that emerges even when using a consistent set of urban design principles, we are able to show that it is not firstly about designing the most affordable or original house type, but about the conscious crafting of the shared spaces. In all our work a sense of enclosure and a human-scaled place has been given prominence, and we have made an effort to imagine and communicate the real experience at ground level through a variety of visualisation techniques.

The principles are not new, but the projects each provided opportunities to develop unique solutions. The project for Kagiso Urban Manage­ment and the Fair Valley Workers Association had to conform to norms of commercial viability which may provide a new benchmark4.


In each project we were given predefined green field sites, which forced us to focus our efforts at this scale. We believe that, for designers, the greater balance of effort has to go into properly defining and developing hierarchies from the scale of the precinct downwards. There is a great deal of challenging design work to be done at the scales between that of the precinct and the front door to each house. The precinct typically incorporates the scale of several urban blocks with associated streets, open spaces and buildings. It is a scale where the spaces between buildings become tangible. Beyond the precinct a further four hierarchies need to be defined and the thresholds and interfaces between them properly ‘crafted’. The hierarchies include public, semi-public, semi-private and private spaces in different arrangements. Urban designers have well-developed skills to define and communicate these hierarchies.

Broader issues, such as how cars and pedestrians access the precinct and how public transport improves overall sustainability, ideally need to be tackled at a different level by strategists in the public sector who have the power to link infrastructural needs to provincial and national budgets and broad political agendas. We do not suggest that city planning is superfluous and unrelated to design, but that it is unrealistic to expect consultants who are commissioned to work on a specific site to have any real influence beyond the site, particularly when there is an urgency to build. In the Fairvalley Project we collaborated with creative private sector planners and transport engineers who communicated with officials and distilled the higher-order fixes. This proved to be most effective in getting us out of the blocks in reasonable time.


Even the most isolated precinct will be subjected to outside forces. The grid and massing need to be arranged in response to these forces, which may include topographical and ecological constraints, logical points of access, alignment of higher-order public transport infrastructure, heritage aspects and legibility. (See Figure 1.) The sooner the set of context-specific constraints are defined and confirmed, the better our ability to respond appropriately and to begin focusing on place-making at a more tangible scale.


With the early introduction of a robust and permeable grid we open up rather than close down design opportunities. Urban designers often refer to the underlying support as an enabling grid. A simple grid is the foundation of necessary affordability, because so much of the cost lies below the ground in the form of service infrastructure. The grid creates a neutral backdrop over which variety can be crafted. (See Figure 2.)

The historic value of a grid system as a support has been well documented in John Habraken’s The Structure of the Ordinary5. His theories show clearly that the grid is not only relevant in advanced societies – it is found in the urban vocabularies of both ancient and developing country societies. It is useful to begin with an efficient, empirically tested grid based on civil engineering criteria and to stop wasting energy and time denying this. The grid is not the place to be original (read funky road layout). The urban blocks, framed by the even grid, represent robust volumes that can be stepped back, carved out and layered in a multitude of ways to create necessary levels of richness and variety at the next level of exploration.


Higher-density housing requires designed, more-or-less continuous edges along the public interfaces, instead of an arbitrary repetition of affordable house types set back from the pavement. (See Figure 3.) Designers need to consider the interface as a continuous, creatively moulded and punctured wall, rather than an opportunity for free architectural expression of each house. Authorship is not important in affordable housing. Place is the client. Without understanding this simple rule, it is impossible to successfully support the in-between as a series of properly defined and appropriately scaled urban rooms that will outlast the whims of the present generation.


Affordable housing and sustainable practice means higher density. For us it has been the single most difficult dimension of the design problem to sell, because compactness is often seen to result in poor environments. In South Africa there is the perception that we remain an exception. People assume a right to greater space because of a perceived abundance of land and a belief that everyone must be given space to express their own individuality. To make the figures work, as well as to build more sustainable total cities, we need to optimise densities and to discriminate against the suburban mindset. Our experience has been that it is often the politicians and officials who frustrate the cause by rejecting density based on their own suburban lifestyle preferences.

If one accepts the challenge to densify, it becomes important to embrace greater compactness of built form while simultaneously maximising the amenity value of the public, semi-public and semi-private spaces. Proper urban space is defined by envelopes into which housing units are set, rather than the reverse of blindly stacking up housing to make up the numbers. (See Figure 4.) Accepting mass as a given and carving the in-between spaces is a very challenging and rewarding part of the design process. It is the skill of balancing solid against void, foreground against background, and urban floor against urban wall. The skill of successfully modulating the in-between takes as much time and practice to develop as an architect needs to develop an understanding of room and facade proportions. Within this process the idea is to create balance, fit, delight and suspense within the whole.


In affordable housing the typologies are consolidated into rows and clusters, typically with shared walls and narrow frontages. It is concerned with the collective rather than the individual. This creates neutral background conditions and continuous streetscapes which are enlivened through creative layering rather than expressive form. (See Figure 5.) The interest is largely created by the modulation of the urban floor to create public courts, pedestrian avenues, woonerven6, parking pockets and so on. Further variety is introduced through considered positioning of larger or more expressive landmark structures against the relative neutral background of houses.


One of the most depressing features of mind­less and repetitive housing types is the lack of extended thresholds, of a series of simple but critical layers between street and front door. This may be in the form of low walls, hedges, trees, stoeps and other rudimentary elements that define intermediary spaces and enliven and soften the street space. (See Figures 6 and 7.)

Before you know it, the money gets sucked into the main structure in order to build more houses and to please politicians. Designers have to fight for it, for the sake of a more dignified total environment. The fall-back option is to allow sufficient threshold space and to conceive the layers as future add-ons. This is however seldom as successful as an integrated, up-front approach. The fact that people seem not to engage in the most simple act of voluntarily planting a tree in front of their houses provides justification for them to be incorporated into the first building increment.


Housing typologies cannot be developed in isolation. Each precinct presents unique opportunities for developing types in the context of very specific place-making criteria that emerge from the grid and street pattern, legibility requirements and different levels of intimacy. Under conditions where densification is encouraged, shared walls, narrow frontages and access to rear gardens usually become influential. In finer-grained settlements the devastating effect of the private car and the turning circles of refuse trucks are soon realised. Urban design offers unique opportunities to demonstrate the negative effects and to argue for public transport connections, shared spaces, pockets of parking and on-street rather than off-street parking.


Settlement by its definition means gradual and negotiated occupation of space. In history this was the usual pattern of development. In modern times it has been substituted by mass delivery, through which there are great risks of excessively limiting choice, thus resulting in sterility. lncrementalism therefore needs to be designed into the system to provide a level of open-endedness and choice in the way people latch onto a robust support framework. (See Figure 8.) lncrementalism linked to a future­end vision allows the opportunity to define the minimum which needs to be built within each increment so that it is independently successful, but even better when ultimately linked to the whole.


1 Second-placed competition entry. Comrie Wilkinson in Association with De Villiers Du Toit. Duncan Village site. Sponsored by the Swedish International Design Agency


2 In association with ACG, Stauch Vorster Architects, Makeka Design Lab, Jakupa and CNdV Africa. Architecture SA Project Award. Urban Design category, 2005.

3 Defined as an intense design exercise typically conducted in two-week follow-up cycles and which includes a selected, manageable group. The group composition may change from week to week depending on the nature of inputs required, while the urban designer is consistently present and holds the pencil, listens and adapts the evolving design. Importantly, the client is also present and begins to understand the value of a ‘thinking on your feet and responding directly to varied inputs’ design approach. Charettes build trust and are both effective and entertaining.

4 Visit fairvalleyecovillage.blogspot.com for a colourful account of the process as recorded by project manager Ori llan.

5 Habraken, J. 1998. The Structure of the Ordinary: Form and Control in the Built Environment. Cambridge: MIT Press.

6 Dutch term for a shared, semi-public space between batteries of houses which contains parking space but is also a communal room for passive recreation.

This text was extracted from an article first published in Architecture South Africa March/April 2009.

View N2 Gateway Project

View Fairvalley Eco Estate Project