We have not known what to do with Battersea Power Station, once voted London's second most popular 20th-century building. At the end of the Eighties, it was to have become a colossal fun-fair, an Alton Towers of the South. Who can remember Margaret Thatcher, when Prime Minister, standing in a hard hat beside one of her favourite businessmen, John Broome - the man who built Alton Towers - announcing the conversion of this temple of power into a phantasmagoric indoor theme park?
The power station was partially demolished, yet no new building work took place. Today, in spite of a change of ownership, the exposed hulk continues to decay. It stands as a metaphor of the underlying weakness of an economy built on dreams of infinite credit and our ambivalent attitude towards such buildings. Do we really want them?
Eastward along the Thames, Bankside Power Station is also redundant - it has been since 1980 - and under threat of demolition by its owner, Nuclear Electric. Like Battersea, Bankside is an impressive hulk, a heroic memorial to Britain's age of industry and power. Its architectural carapace (the work of Sir Giles Gilbert Scott, who also dressed Battersea) is a thing of brute beauty. Its sin, in its detractors' eyes, however, is not so much its architecture as its bedevilled history.
Planned in 1947 and not completed until 1963, Bankside does the unthinkable: it challenges the authority of St Paul's Cathedral. Perhaps it was a planning blunder to have it built here opposite Wren. Whatever, the sin was compounded because its chimney was limited to 325 feet in height (to be lower than the dome of St Paul's). Thus constricted, the smoke from its oil-burning boilers wafted over the river and added to the layers of filth that had changed the colour of Wren's temple from white to black.
Despite three attempts by English Heritage to get Bankside listed, the Department of Environment has declared it 'immune' from such treatment until 1998. Nuclear Electric, its architectural adviser Sir Andrew Derbyshire and the London Borough of Southwark, all want to see an end to this brick colossus as soon as possible.
The building has one realistic hope of survival: it could become the home of London's Museum of Modern Art. That decision lies with the museum's proposer, Nicholas Serota, director of the Tate Gallery. He is considering three sites, but must make up his mind over Bankside before Christmas.
Should he decide against it, the building will be under very real threat of instant demolition. This would be sad, not only because it could be converted into something useful, but because what will replace it - a mixed bag of offices and houses - is unlikely to contribute anything positive in townscape terms to this stretch of the Thames.
This is not to say that Mr Serota should be under pressure to save Bankside. There are alternative sites for the Museum of Modern Art, and the question of whether to design an entirely new building or to remodel an existing one cannot be taken lightly. When the New York Museum of Modern Art was conceived 60 years ago, there was little question as to whether or not it should be housed in a Modern building. It was, and Moma remains an exciting and enjoyable place to visit and work in. A purpose-built design for London's Moma would demonstrate confidence in contemporary design and show that the new gallery was a fresh and vital venture.
Some of the world's best contemporary architects are, however, in the process of transforming redundant temples of industrial power into world-class museums. In Turin, Renzo Piano is at work on a long-term project to turn the vast Fiat plant at Lingotto on the edge of the city centre into a complex cultural centre. This is the famous factory with a racetrack on its roof. In the history of 20th-century architecture, Lingotto plays a heroic role; it could not have been demolished without great physical effort - it is on the scale of a small town - or without robbing the world of a wonder as marvellous as the Eiffel Tower or Brooklyn Bridge.
In Germany, Sir Norman Foster and Partners have drawn up plans for the conversion of a titanic boiler house at a redundant mine at Zollverein, Essen. A design centre and museum will be housed inside the building's five boilers. The memory of what the building was and the grimy but essential tasks it shouldered will be a part of the new museum's fabric.
Although architects of this calibre could perform wonders with Bankside and it could be linked by a new footbridge to St Paul's (and the 3 million tourists who visit it each year), there remains the underlying problem of acceptance. Where the Lingotto factory played a key episode in the history of architecture, materials and structural engineering, Bankside has been no more than a backwater. Where the local authority and community support - and finance - the Foster scheme in Essen, Bankside is under threat from local politicians.
It is not, sadly, a building that makes friends easily. When plans were announced to turn the Gare d'Orsay in Paris into a museum (now the Musee d'Orsay), they received almost universal support. The station was designed to be festive; it is a flirty structure and its style appeals to rabid fogy and zealous Modern alike. Bankside, by contrast, is an acquired taste.
Yet Bankside, unfairly, has been getting a bad press. In this week's Sunday Telegraph, Sir Andrew Derbyshire claimed that it could not be converted at all easily into an art gallery because the machinery installed inside could not be removed without at least taking the roof off. 'Nonsense,' says Alan Baxter, the engineer whose proposals for the future of Bankside have been seen by Mr Serota. 'In strictly engineering terms, what went in can come out. It's simply not a problem.'
The Tate has ventured into the remodelling of industrial buildings once before: the Tate of the North was worked into Jesse Hartley's magnificent Classical warehouses at Albert Dock, Liverpool, by James Stirling and Michael Wilford. This was a popular move. Will it work again at Bankside? If the right architect is chosen, if the right bridge can be built across the Thames, if the money adds up.
Housing the Museum of Modern Art at Bankside would be a popular move among conservation circles. It offers a massive amount of ready- built covered space with no interiors to upset. It might also attract funds more readily than a brand- new building that - unless it was severely compromised like the timid extension to the National Gallery - would almost inevitably be unpopular in Britain for the first 30 years of its life.
To date, the Tate's buildings have been conservative. The parent gallery is a housed in a rather limp Roman temple designed by Sidney Smith, begun in 1897 and fiddled with over the years (the impressive sculpture galleries and central cupola by J Russell Pope date from 1937). Additions by Llewelyn Davies Weeks in the Seventies and by James Stirling (completed 1985) are not modern architecture at its best or most inventive. The Tate St Ives, opened this year, is as cautious and crustaceous as a hermit crab, while Liverpool is a veteran building intelligently reworked.
Before he went to the Tate, Mr Serota added coolly and sensibly to the Whitechapel Art Gallery (Charles Harrison Townsend, 1897-99; alterations by Colquhoun and Miller, 1984). Perhaps, now, it is time to let rip and commission a brand- new building free from the convolutions of conservation politics and the weight of precedent.
This could be the right decision for the Museum of Modern Art and the wrong one for Bankside. But no redundant building - as Battersea Power Station proves - can rely on one project to save it. As a museum of science or architecture, as an indoor sports centre, a complex of theatres and galleries, Bankside might be perfect. As a Museum of Modern Art, it would be a very English compromise, like St Paul's itself. No bad thing? The decision for or against Bankside is as big as the building itself.