Web 2.0 is largely built in the form of third party hosted applications. Users "rent" (in the technical sense of the term) access to applications, and the seller has exclusive control over what those applications will do in the future.

That leads to plenty of ills and market inefficiencies. Users become locked into various services. If we make the simple assumption that users' needs will diverge over time, it follows that not only do they easily become locked in, but then many of them are certain to be poorly served. Also, privacy issues are substantially created by the centralized service architecture of Web 2.0.

Is there an alternative?

The Personal Computer revolution was the first time in our industry when a mass market of consumers could have complete control over a computer: choosing exactly what software to install or not install; what to run or not run.

The Web deserves an equivalent moment. The isolated PC, in an office or home, switched off at night, poorly connected to the Internet is no longer of interest. The Web needs Personal Web Services: web servers that a mass market of consumers can acquire and fully operate.

The virtualization and commoditization of web services is already under way. For example, virtual GNU/Linux instances can be leased on load-driven demand for relatively little money.

Such projects, however, are likely permanently too low level to reach a mass consumer market. The degree of complexity of managing such a server is simply too high, and likely "untamable".

Perhaps more importantly, as inexpensive as virtualized GNU/linux servers are, they are still only economical under relatively high loads: they are optimized to be work-engines in high-demand environments. A truly personal web service should be able to idle quite efficiently and to share the cycles of a physical machine with quite high granularity (so that many instances can be concurrently run, most of which are idle).

Flower hits a sweet spot. It is slight enough and optimizable enough to run very inexpensive instances of it, yet it is also powerful enough to write complete server-side applications. Flower instances are a good candidate for the Personal Web Service.

There is another aspect to Web 2.0: the benefits of data aggregation; the wisdom of crowds as content.

If people used nothing but their own private circles, would those benefits be lost? Hopefully not.