The final list will go to federal cabinet in March, but much work is needed before then, as most projects submitted were poorly defined with weak economic analysis, did not look at alternatives and rarely considered land use. Sydney's new plan for land use around stations is thus welcome.

I say most cities have serious rail projects, because Perth does not. It has spent the last 20 years modernising and extending its railway, largely because of the political backlash to a conservative state government's attempt to close the railway down - in the middle of the first major oil crisis. It mobilised everyone and showed how important a modern electric railway was to the future resilience of the city.

Perth's new Southern Railway already is something of a model for other cities as they prepare their plans. It goes deep into the outer suburbs, places badly hit by last year's oil crisis (a foretaste of many more). Now, 50,000 people a day are carried along a corridor that before could only manage 14,000 on buses.

It cost $30 million a kilometre, including a one kilometre tunnel under the city and two river crossings. Sydney's recent Epping-Chatswood rail project cost $200 million a kilometre due to the extensive tunnelling.

The Southern Railway is a model in other respects, too. Most significantly for Sydney, some stations have transit oriented developments, or TODs. These are high density, mixed-use areas with homes, shops and offices, designed to integrate closely into the station and which can provide a local area with many city services.

In the US such developments can allow residents to reduce their car usage by 50 per cent and increase their household income by 20 per cent. They are popular therefore with those who can afford to live there (affordable housing can be built; Vancouver for instance, requires 15 per cent of each such development to be devoted to affordable housing).

They are popular with councils, as residents spend much more money locally from the income they save, and from a government perspective they are an obvious way ahead. They can help fund the rail project (as was the station complex at Chatswood) but they will save considerably on the cost of serving new development on the urban fringe.

In one study we did for Parsons Brinckerhoff we found such developments saved the government $85,000 per block on urban infrastructure. Over 50 years, they would save $250,000 per household in transport costs, let alone the greenhouse gas and health savings.

High-density developments around railway stations are not new to Sydney, but the notion that dense centres will be built in the suburbs always produces a reaction from any local community.

We can learn much to address such concerns from Vancouver, a city that has built more TODs than most. It requires a small percentage of the value of any new development to be spent on local social infrastructure - landscaping, cycleways, child-care centres, even an art house cinema in one case - to be determined by the local community. Everyone gains in these development processes.

In Sydney new dense TODs should have a special zoning regime, with a set of guidelines that ensure densities are optimised, parking is minimised, affordable housing is included, green buildings are required and that a local process of creating social infrastructure is an essential feature.

The revival of rail across Australian cities will be an important phenomenon, but to make the most of it, it is essential that the projects cluster high-density developments around the stations.

Peter Newman is the Professor of Sustainability at Curtin University in Perth and a member of Infrastructure Australia's Advisory Council.