I am convinced that many problems with Bitcoin would be solved if we could come up with a way to tie mining difficulty to the currency's demand instead of only to the mining computing power. Instead of ensuring a predictable generation rate, this could ensure a predictable <i>value</i>. When demand wanes, the currency is harder to mine and becomes more scarce, and when demand surges, the currency becomes easier to mine and more available. This would stop some of the volatility in BTC prices. You could plan the rough change in value over time, and build it into the algorithm, perhaps making gains unlimited at first, then making it very deflationary for several years, and gradually easing gains to something like 0-2% deflation per year. A currency with a known long term future value is obviously very useful, and a stable value is also more useful.<p>I can't think of a decentralized way to measure demand, though, so maybe this is not possible.