What I find a very important distinction, and one this article quite completely misses, is whether it takes a great deal of effort to revive a project that has been abandoned by its community. A project is easy to revive if it still has its source code easily available, and the source code is in a condition that makes it easy to build; at that condition, anyone can pick a project and start using it - and usage often eventually leads to development.
In this way, there are _really_ many projects which are lingering, sleeping, or just waiting for something to come up. When was the last update to GNU m4? Does somebody wait for the next release of Knuth's implementation of TeX? Numerous small applications (I have in mind e.g. PikiPiki, the Python-based simple wiki implementation) just lie there, available on the web, but occasionally taken into use and built upon. Many projects never had much of a community, but still manage to be useful, indispensable even, for a group of users that know little about each other.
This "long lingering" is one possibility that IMHO really makes open source different from commercial products. *If* a company really can take a working product away, then its future is _always_ uncertain in a way that truly open source projects never face. *If* a product can be built upon by anyone that has the motivation to do so, it is really hard to kill it completely.