Dog legs

The urban ‘dog leg’ is an under-used arrangement, I reckon. Here’s a nice one in Durand Gardens, Stockwell: note the house wedged into the angle.

It’s a delight to walk around this corner: there is some rat running but the acute change in direction does seem to deter drivers. And the people at Space Syntax insist that their urban planning ‘rules’ can be OK with the occasional dog leg, even though they restrict sightlines. My ideal design for green belt (or, I suppose, suburban) housing (see these earlier posts) also features dog legs:

This is deliberate: it’s a way of making a ‘lazy grid’. This in turn supports a subtle hierarchy of streets, where straighter roads are (slightly) higher in the hierarchy than more meandering routes. (Tree planting (or absence of), street width and presence of public transport route also factor into the hierarchy.) The dog leg inflection is good for siting an entrance, typically in this design to the lobby of a ‘multi-family’ building where a degree of footfall can be expected. Below is a sketch of one of these conditions.

This dog leg isn’t deep enough to completely block through visibility, and everything is well surveilled: there are no blank ‘end of terrace’ frontages, as deprecated in e.g. Secured by Design. The geometry here is all orthogonal and may seem quite rigid (then again, look at Mayfair’s estates) but this is an ideal design and it’s assumed that a real site, with real topography, existing streets to connect to, extend, etc., would cause things to move around a bit.


Building in the green belt: 3

One archetype of desirable urban life is found in the urban village. This conception of place in the city owes much to Jane Jacobs but is also seen in i.e. Abercrombie’s 1943 analysis of London’s neighbourhoods. The urban village typically has two faces: the high street face, and the ‘countryside’, or park face. I argue that both are fundamental. Ideally, both are found close together: they allow the village to be understood and enjoyed as a village. On the one side, shops, eateries, small business, primarily indoor social gathering spaces. On the other side, green open space, room for exercise and outdoor socialising, play, escape. In between: homes (and institutional support, i.e. schools). In the urban village, everything local of value is walkable.

Northcote Road, Battersea, lies at the centre of an attractive and popular urban village. A quick look at the street layout of this neighbourhood shows that it closely tracks the model described above. Northcote Road itself is the high street. The countryside face of this village is – jointly – Wandsworth Common and Clapham Common, two decently sized urban parks. The perpendicular streets in between – Wakehurst Road for example – represent some of the most desirable streets in London for dwellings of this type and scale.

The block highlighted above is 270 m by 60 m; dimensions which – along with the simple grid topology – permit the walkability of the neighbourhood. In addition, Northcote Road, although narrow, carries buses for most of its length; it is a Connector that becomes a High Street (in Transport for London’s street hierarchy), with both movement and ‘place’.

Similarly, the idealised green belt typology which I sketched out in the previous post (here) takes its place in an ‘urban village’; this time, a green belt village. Although by definition not a central district, it is not treated as suburban: it remains urban. It makes up a neighbourhood with a similar topology and scale to Northcote Road, Battersea. There is a ‘lazy’ grid topology with blocks with a slightly smaller perimeter (170 m x 95 m).

Trams are routed along the central connector, which generates enough footfall and ‘place’ to support commercial spaces. This street could take on the specific character of a street such as Northcote Road; however, other precedents are available. For instance, shown below is a view of a main road in the Vauban district of Freiburg:

The character of the high street in my ideal model neighbourhood is probably closer to Vauban than Battersea. As the sketch below shows it is wider, at 36 m, and contains pavement seating, pavement, segregated cycle lanes (on both sides of the street), and a roadway for light vehicular traffic and trams. (Trees from the intersecting local roads are visible, but the high street is not planted as an avenue as this is bad for commerce.)

What really matters here is the basic hierarchy of connector to grid to green space. If everything of value in the neighbourhood is to be walkable, there must be quick access to not only amenities such as high street shops but also to transportation for non-local journeys. The planning principle followed in the model is to keep foot journeys to under 800 m at a maximum. Below is the furthest possible foot journey form dwelling to tram stop, at 630 m:

The reverse of this journey represents the furthest distance from dwelling to park. The model is also car-free, with exceptions made for pick-up and deliveries: this means that the streets can be driven but the only parking provided is in a neighbourhood multi-storey parking structure. Again, this is easily reached on foot; the route shown below is 500 m:

In the next (and probably final) post in this series, I’ll look at the sorts of planning principles that can bring additional variety and character to a new neighbourhood that is set out functionally along the lines described here.

Building in the green belt: 2

The challenge of building in the green belt is typological: we may know that we want to live there, but in what sort of dwelling? The promise of living at the city edge is space: traditionally this has meant houses; detached or semi-detached, each with a private garden. The price of this model is low density and its corollaries:  inefficient land use, hard-to-reach diffuse public transport, car dependency and its associated health impacts. But if higher density is the answer, what is the best higher density model? Terraced housing offers many advantages. Density can be moderately high. Every dwelling has identity and presence; a front door at street level. Every dwelling can have a garden (or two; one at front, one at back).

Typically, terrace back gardens remain fenced off from each other; however, in some places you see recognition of the potential of a connecting space within a terrace block interior: the image above shows a planted connecting passage in Middlesborough. Similar examples are found in Japan: Haruki Murakami describes such a space in his novel The Wind-Up Bird Chronicle. I think these spaces have further potential. The massing below describes a hybrid typology of terraced houses and point block towers (although slender, at one dwelling per floor). Together, these types enclose a block whose interior features a large, shared but secure garden for residents of all age groups.

Below is a part plan at ground level:

The shared garden is fringed by the private gardens of terrace houses but is also overlooked by the point block dwellings; for these blocks, the shared garden offers a park-like setting in the style of post-war modern housing, only in miniature, and without the radical ground plane permeability of that era: in effect, the proposal also hybridises models of land enclosure; some is private, some is shared but only for residents, some is public (street; pocket park).

Some figures. The plot is 1.3 ha; the built area is 3,545 m2, or 27% of the plot; a relatively low figure. Interior gardens, private and shared together, are 33.5% of the plot; the shared garden is 15%. Front gardens, footpaths and public open spaces together take 27% of the plot. The density is 274 hr/ha, and apartment dwellings provide 13.5% of the habitable room count; most residents here live in houses.

If most residents live in houses, and this type is doing most of the ‘work’ in terms of density, why provide towers at all? One reason is that household sizes vary, and it seems odd to confine smaller households to city centres; why shouldn’t single people or couples also live ‘in the countryside’? At the same time, one or two bedroom ‘houses’ are a difficult typology; these are implicitly one or two storey dwellings with diminished presence on a street and lower density. Duplex stacking offers a partial answer but the advantages of flatted development – with the right plan – are light and views. The freehand perspective above shows how the point blocks in the hybrid massing take full advantage of long views within the block, as well as having sufficient height to give distant views. Each floor in one of these blocks is a single ‘360 degree’ two bedroom dwelling, as shown in the plan below.

The roofs of the point blocks are planned as resident gardens, screened by the facades. A similar approach is taken with the houses to the corners of the block. Here, the interior angle makes it impossible to plan a garden at ground level for every dwelling; instead, the roofs of the houses are used as gardens as shown below.

You might have spotted that no car parking is provided. In the next post, I’ll sketch an approach to urban design, street layouts and transport that can complement this typology.

Building in the green belt: 1

This is a post in two parts. In this first part, I’ll set out some reasons for considering green belt development. In the second part, which will follow (eventually), I’ll describe a test case proposal for a specific green belt site.


People have made good arguments for developing some of London’s green belt as housing. After thinking about this for a bit, I agree. The quantity of land designated as green belt is excessive, given the intense pressure on London’s existing housing and the escalation of house prices. Green belt designation helps to create that pressure and the rise in prices: it is inequitable policy.

I’m not arguing for the abolition of green belt: I think it should be reformed. Not all green belt land is of equal quality. This helps enormously when it comes to considering how some of it could be converted to housing. We are not talking about national parks, or anything much resembling ’nature’. Instead, large tracts of green belt are used as monocultural farmland (leading to the argument that suburban housing, with extensive hedge planting and gardens, delivers better biodiversity). Big chunks are laid out as golf courses. Some green belt is not even green, and of non-existent amenity value: gravel pits, surface car parking.

Still, there are people who believe that the green belt shouldn’t be touched. What are their arguments?

1: Limiting the city

One argument for retaining the green belt is that limiting the city is good in itself: why should the city be allowed to grow indefinitely? This connects with arguments about immigration: why should the population grow indefinitely? Here I think we run into a kind of fallacy where principle is argued against degree. Most would accept that at some point a city, or a region, or a nation, is ‘full’ and should not grow further. The issue is whether or not we’re at that point. I don’t think London is ‘full’. I also don’t think that the very high densities typically now proposed for urban sites are acceptable, or that existing lower density areas can be made denser to a sufficient degree (I will discuss this further below). Therefore, to allow for growth, the green belt constraint on the city should be relaxed. Exactly how far is a matter for further argument, but it’s an argument in which all sides can accept the principle that there should be a limit to growth. Advocates of green belt development are not saying that the city should grow forever, until everything is ‘covered in concrete’. Instead they tend to agree, as I do, that there should be a green belt; just not the green belt that we have.

2: Non-green belt solutions

Another argument accepts that the green belt contributes to the pressure on housing but claims that this pressure can be relieved in other ways. Development of the green belt is therefore unnecessary, at least for now. This is a much better argument. But let’s look at how urban intensification is currently done.

Open market sites

The key fact here is that suitable open market urban sites in London are scarce and expensive. This has led to a very large increase in density in those places where development is happening. The London Plan calls for a maximum residential density, for sites with the best access to public transport, of 1,100 habitable rooms per hectare. A typical modern housing development will tend to exceed this density by a large factor (2,000-3,000 habitable rooms per hectare is not uncommon). This has two consequences. One is height: London is rapidly acquiring towers, or high slab blocks. These are relatively expensive to build and carry significant environmental penalties: poor daylight, wind at street level, and visual impact on the skyline. Another is intense pressure on the amenities that go with housing: child play space, car parking, schools. Partly as a result, developers have aimed this style of development at people without families: that is, occupants who can accept relatively small dwellings, don’t need cars, and don’t call on facilities such as schools or parks.

Non-traditional urban sites

In response to tower proliferation some people have called for moderated density, albeit still an increase, and have looked for alternative urban sites across which non-high rise housing might be distributed. A recent competition organised by New London Architecture threw up a range of ideas. In addition to (the usual?) capsules and houseboats we saw residential schemes that in effect double count land: housing on top of retail car parks, on top of schools and hospitals, on top of railway lines, and in one perverse proposal, instead of roads (refuse collection and fire brigade access to be determined). But the thing to note here is that the principle of mixed use development already informs planning policy in every London borough and has not significantly moderated the cost of land or reduced pressure on housing. Stacking uses is hard. The 2005 Gerrard’s Cross tunnel collapse shows the risk of building over a working railway line. It is almost impossible to imagine building on a working hospital, or even on top of a school. How do you connect services? How do you guarantee un-interruption of existing services? (Hospitals are very highly serviced.) How do you even get to your front door? (Hospitals are secure buildings, and schools require secure sites.) Housing on top of shops is feasible if starting from scratch: this was first seen on a large scale in the 1990s with the development of the Cromwell Road Tesco, and is a feature of the recent Woolwich Central development (and perhaps undeserving winner of the 2014 Carbuncle Cup). Large scale stacked use schemes may have a place but raise liveability questions and are difficult to construct: I don’t foresee them arriving in quantity.

Suburban intensification

More promising is the idea of suburban intensification. Much of suburban London is low density: it is largely under-utilised land. But simply observing that large areas of London are low density is not enough. It is true that, were these areas to be rebuilt at the density of Islington or Chelsea, there would be a large increase in housing stock, and without need for high rise. But these areas cannot easily be rebuilt wholesale: the land is already divided and owned. Hence one line of investigation is to see if existing households could be helped or incentivised to sell, for higher density replacement development. But even taking an optimistic view, land release here is likely to be slow.

Council estate redevelopment

Perhaps most promising of all is redevelopment of low density council estates (and some are low density overall despite inclusion of high rise). This type of redevelopment is already happening in several central London boroughs (I live in the recently redeveloped Myatt’s Field North, Lambeth), and the model could be extended to outer boroughs. It is politically sensitive, and tends to be described as ‘social cleansing’ if existing council homes of similar quantity and quality are not re-provided in the new development. There is a risk that sites that are made available in this way will be unpopular with purchasers, who do not always want council tenant neighbours; this, in turn, pressures developers to reduce or exclude the social housing component. And again the land is not empty, it is built; hence there is sometimes a risk to significant existing architecture, such as Robin Hood Gardens in Poplar, East London, which is to be demolished as part of an estate intensification scheme. This is a conservation issue, which should be weighed against the desire to retain green belt: is it really better to sacrifice a good example of modern housing over even the lowest quality green belt?

Council estate redevelopment at higher density – combined with whatever can be gained from suburban intensification – can help to grow the city while leaving the green belt untouched. But is it enough? And are advocates of this route to growth adopting a balanced, moderate, or fair position? When combined with fastidious preservation of green belt, advocacy of estate redevelopment looks partisan. Advocates of this route only to growing the housing supply are open to the accusation that they favour one constituency, those who enjoy green belt adjacency, over others; those who live in and around council estates, and those who live in suburbs targeted for densification.

3: Escape from urban life

A final argument for leaving the green belt untouched is that it provides essential amenity – an escape from urban life. I agree that some of it does: I spend plenty of my weekend days biking up and down the Surrey lanes popular with cyclists. However, a point made by Howard – the originator of the garden city idea – still stands: the ‘countryside’ is often not accessible. It has open spaces but also ‘no trespassing’ signs. Knowing that much green belt is farmland, this shouldn’t be surprising: farmland is rural industry; someone’s productive resource, not a leisure resource. And then there are the golf courses and country clubs; also typically enclosed, with no right of access. There is also the issue of infrastructural barriers to access: the green space exists, and perhaps you have the right to go there, but can you as a practical matter get there?

Howard answered both of these challenges with one idea: the garden city would enfold the green, developing some as housing and employment and making the rest more accessible than before, both as a matter of right, and physically. Howard is often taken for an advocate of the suburb but he envisaged small, relatively dense and delimited settlements, interspersed with a mix of public parks and farmland, and linked by fast rail connections. Compactness produces practical variety: if you live two hundred meters from a green space, you are likely to be able to go there often. Rail connections give further choice: if you are bored with the settlement where you live, and with its amenity offer, you can quickly travel to another. His emphasis on rail as a key enabling technology is in part down to context – cars existed only as prototypes when he wrote Tomorrow (published 1898) – but increasingly, rail – efficient, sustainable – is vindicated as the transport mode of choice. But new rail links are only justified when they serve development. Counter-intuitively, development – by bringing the railway – helps with the escape from urban life.


There is huge demand for housing in London. Although the green belt brings certain benefits – principally leisure – through restricting supply it also harms. As it stands, green belt legislation is bad policy. It preserves open land of questionable quality. It preserves land that is often hard to access, undermining the claim that what is being preserved contributes to London’s amenity. Efforts to find non-green belt sites for increasing housing supply not only risk falling short, they risk leading to extremes, such as planning to redevelop most of London’s council estates. That we are contemplating such extreme measures only highlights the unfairness of the institution of the green belt. It deserves reform.