A boundary is not a verdict.

Which side of the line the site is on.

National policy now splits the presumption in favour of sustainable development in two. Inside a settlement, a proposal should be approved unless the benefits are substantially outweighed by the adverse effects. Outside one, only certain forms of development should be approved at all, and housing has to find its way onto that list. So the first question on a rural site is never the design. It is whether the site is inside or outside, and what the development plan says the boundary is.

A settlement means cities, towns, villages and other predominantly built-up areas, and it expressly includes land which is allocated or already has permission and will form part of the built-up area once it is finished. That last point is worth money: an allocation or a consent next door can move the line. Hamlets and scattered groups of houses outside built-up areas are not settlements unless the plan says so, and villages washed over by the Green Belt are excluded from the definition entirely.

The edge-of-settlement case has its own rule. Where a site falls partly inside and partly outside, both policies are applied to the relevant parts before the decision maker comes to an overall view. A site plan and a proposals map, read together, will usually tell you which conversation you are about to have.

Outside the boundary is a list, not a no.

The list is finite and it is written down. For housing the useful limbs are the reuse, extension or replacement of an existing building of permanent and substantial construction; the redevelopment of previously developed land, including a change of use to residential; limited infilling within groups of houses; a rural exception site for affordable housing, or community-led development; land already allocated in the plan; and development which would address an evidenced unmet need.

That last one is where most edge-of-settlement housing lives, and it is stated in policy rather than left to a balancing exercise. An unmet need expressly includes housing where the authority cannot demonstrate a five year supply of deliverable sites, or scores below 75 per cent on the Housing Delivery Test. The scheme then has to be physically well related to an existing settlement and of a scale the infrastructure can take. Those are the two limbs to design toward, and they are design questions as much as planning ones.

One carve-out matters. An isolated home, meaning one lying outside settlements or groups of houses, is not on that list at all. It sits under a policy of its own, the bar is much higher, and very few schemes clear it.

Green Belt is a different word, and a harder test.

Green Belt policy sits outside the settlements framework and runs on its own terms. Development in the Green Belt is inappropriate unless it falls into a defined category, and inappropriate development is by definition harmful. It should not be approved except in very special circumstances, which do not exist unless the harm by reason of inappropriateness, plus any other harm, is clearly outweighed by other considerations. Substantial weight is given to the harm, including harm to openness.

The categories that are not inappropriate are the ones worth knowing. They include the reuse, extension, alteration or replacement of an existing building, provided it is of permanent and substantial construction and lawful. A replacement must be for the same use and not materially larger than the one it replaces. Extensions are measured against the original building, which means the building as it stood in 1948 or as first built if it came later, not as it stands today. They also include limited infilling in villages, affordable housing for local need, and the redevelopment of previously developed land.

Grey belt is the newer route and it is worth reading closely. It means previously developed land in the Green Belt, or other Green Belt land that does not strongly contribute to three of the five purposes the designation serves: checking the unrestricted sprawl of large built-up areas, preventing neighbouring towns from merging, and preserving the setting and special character of historic towns. Safeguarding the countryside from encroachment, the purpose that open Green Belt land most obviously serves, is not part of that test. Development on grey belt also has to answer an evidenced unmet need, sit in a sustainable location, and, if it is major housing, meet the Golden Rules: affordable housing, the necessary infrastructure, and new or improved green space the public can reach. Until a plan sets its own figure the affordable requirement is 15 percentage points above whatever would otherwise apply, capped at 50 per cent, and 50 per cent by default where no requirement exists at all. That is a viability question before it is a design one.

The White House near Hatton is what that argument looks like in practice. A tired 1950s bungalow and a detached garage, 230.9 square metres between them, washed over by the West Midlands Green Belt and outside any village boundary. The replacement was consented at 242 square metres, an increase of 4.8 per cent. The council reached 22.9 per cent by excluding the garage it had itself counted at pre-application. A Planning Inspector allowed the appeal in March 2021 on a plain comparison of the new house against the one it replaced, finding the increase in volume modest given a single storey design split into blocks by a glazed corridor.

In a National Landscape, scale is the argument.

In a Protected Landscape, meaning a National Park, the Broads or a National Landscape, substantial weight is placed on conserving and enhancing natural beauty. Proposals are expected to be limited in scale and extent, and sensitively located and designed to avoid harm to the statutory purposes and special qualities of the designation. Major development should be refused other than in exceptional circumstances, and where it can be demonstrated to be in the public interest.

The word that decides a scheme is major. It is not a floorspace figure or a unit count here. Whether a proposal is major development for this purpose is a matter for the decision maker, taking into account its nature, scale and setting, and whether it could have a significant adverse impact on the purposes the area was designated for. That is an argument you can win or lose on the drawings, which is why the layout and the amount of land you agree not to build on are planning evidence and not just good manners. Sites outside a Protected Landscape but within its setting carry their own, lighter test.

At Oakridge Lynch the whole site sat inside the Cotswolds National Landscape, on Cotswolds Steep Slopes, and just outside the settlement development limits. Nine homes were consented on the northern half. The southern half was kept as open meadow rather than built out, and the houses were held low and loose against the village edge, in natural stone with gabled roofs, so the row reads as a village edge rather than an estate. Design was one of the eight grounds the council refused it on. On appeal the scheme was found to integrate well with its surroundings, with no conflict with the design policy.

What the village asks for is evidence.

Rural applications are decided in a small arena. There is usually no allocation, no design review panel and not much on the file beyond the application itself, so the local material carries more weight than it would in a town. A parish council that has written down what it would support, and a housing needs survey commissioned by that parish, are evidence in a way that a consultation exercise run by an applicant is not.

At Butlers Marston, a village in south Warwickshire where the primary school closed in 1996 and the site has stood empty since, the parish council set out its terms in writing in January 2026. Additional parking, for the parish, to relieve a nearby lane. Up to nine properties. A mixture of materials, stone on two sides and brick on the others. Site traffic taken off the main road rather than through that lane. Footpaths through the site linking to the paths already there. Houses set back from the road with greenery in front of them. The application that went in answers each of those.

The parish had already asked its own residents. A housing needs survey delivered to every home in the parish in 2022 drew 43 responses, a 43 per cent response rate, which the report itself calls excellent for a survey of this type. Asked what should happen on the former school site, most respondents said six to ten properties, and most wanted a mixture of types and tenures. A public meeting in January 2026 was attended by about 33 local residents, and a show of hands indicated that a majority of those present were broadly supportive of the principle, subject to the matters they raised. Five revisions of the scheme were presented to the parish before anything was submitted. The application is with Stratford-on-Avon District Council and has not been determined.