I am not sure about the statistics, but it seems that at least 4 out of 5 commercial software projects (or for that matter open source ones) fail to achieve any sort of success. The statistics is likely to be much worse. The result is that the code written for these projects get effectively forgotten and lost. While the project/product as a whole might not have been a success, there are often small modules in these projects that are real gems. I will go so far as to say that some of these gems are good enough to be launched as separate products. Just because of the fact that these gems are part of a larger mess, they often do not get to see the light of the day.
If however, your modules are really loosely coupled with the rest of the code, you can actually re-purpose the module for something else. And as I said, a small and insignificant looking module can actually be turned into a product if it is good enough. While the problem of hidden gems is not that grave in the open source world, it is a significant problem in commercial development setups. We need to start writing highly separable components and mine these even if the project as a whole gets canned.
