Learning from mistakes, not repeating them
The policy skills of the NSW Labor government were graphically illustrated by Kristina Keneally’s recent announcement that the scheme offering a 60c/kWh feed-in tariff to anyone who installed solar panels on their roof had attracted far more demand than the government had budgeted for, and that the price would therefore be cut back to 20c/kWh (roughly the delivered price of coal-fired electricity). This pattern has been repeated with a string of schemes in Australia and overseas in recent years. The reason is simple, as can be seen from the graph below.
The retail price of solar modules has fallen by about 30 per cent in the past few years, after several years of stability as the industry developed its own sources of silicon supply. (Before about 2003, the silicon needs of the solar industry were supplied by cheap offcuts from the [then] much larger semiconductor industry.) Add to that the appreciation of the Australian dollar and you have a huge cost reduction, so that an offer that seemed marginally appealing when the plans were being developed rapidly became a no-brainer.
There’s every reason to expect this trend to continue. Despite fluctuations in particular countries caused by episodes like this one, demand is growing rapidly, and bringing with it scale economies and induced innovation at every stage in the production process. Annual solar PV installations have risen from around 1 GW (capacity) in the early 2000s, to an estimated 10GW for 2010, and growth remains rapid. 
On the other hand, at current costs, home-based solar PV is still a high-cost option. My suggestion would be to estimate the likely price of electricity under the kind of carbon price needed to achieve a serious reduction in emissions, calculated the implied electricity price for coal-fired power, and set the solar feed-in price at that level. A rough estimate would be: 20c kWh likely price without a carbon price, $100/tonne CO2 = 10c/kWh, so feed-in price = 30c/kWh. At that price, early adopters would be paying more than the cost of the grid option, so only those willing to make some sacrifice would sign up. From the viewpoint of suppliers, the price as I’ve calculated it is the relevant long-run marginal cost, so it’s just a downpayment on a comprehensive carbon price.
fn1. For comparison, new wind installations were 37.5 GW capacity in 2009, nuclear additions for 2010 so far have totalled about 4GW, and construction initiation totalled about 10 GW – source IAEA. Taking account of the higher capacity factor for nuclear, I’d estimate the net additions of nuclear and solar PV were about equal. Considering the likely construction times for nuclear plants, and observed growth for solar, it seems likely that solar will be a larger net source than nuclear over the next decade. However, I don’t want yet another slanging match on this topic. All discussion of nuclear, pro or anti, should go to the relevant sandpit. I’ll open another if necessary.