How do you assess such a place? As architecture, as housing (often a very different thing), as an industrial process? It has to be the first. Because it really shouldn't matter one jot how a building is put together, or what its use is, so long as it works and it looks good. All you need to know is that the architect set out to achieve something, and achieved it. In this case, the architects have plenty of good track record. Allford Hall Monaghan Morris are one of the more successful young firms of recent years. They've done good schools, they've done fine offices, they are working on humanising the Barbican arts complex in London, they have even made a bus station in Walsall look interesting. They are now at that point, which comes to all aspirational architects, when they have to publish a big thick book with an academic gloss about their work, just out. Yup, they've arrived.

Simon Allford is the one in charge of Raines Court (named after the old Raines Dairy that used to stand on the site, sandwiched between older housing blocks and a railway line). Allford is entirely candid about the drawbacks of the prefab approach. It's great in principle, because you can make buildings in controlled conditions in factories rather than out on messy, dangerous, weather-dependent building sites. But the trouble is it's not cheap enough, or prefabricated enough. It's not really Lego enough, really. When I meet him at the site - glittering in the sun, excited prospective tenants being shown round the surprisingly sizeable flats with their commendable absence of corridors - Allford remarks stoically that while he could order the big steel-framed boxes for the rooms easily enough, and get them stacked them up in double-quick time to make his building, he then had to laboriously fix the outside skin, the access walkways and so on. Ideally this wouldn't happen. Ideally everything would arrive in a finished state, plug together and that would be it. It's not a big enough industry yet, this prefab lark, and it's still too craft-based, too labour-intensive. For this reason the Peabody Trust - which has now done three experimental prefab schemes, and will do more - is finding that, paradoxically, it's an expensive way to build. It shouldn't be, but it is.