I keep seeing articles and quotes where people in total ignorance state that Open Source is dying and then they turn and point at the number of dead projects on SourceForge to back up their ridiculous claims.
Quick note, I'm going to use the term "Open Source" in that silly loosey-goosey way many "journalists" in the media use it. Otherwise this mini-article looses its oomph while I do vocabulary translation.
There are a lot of dead projects on SourceForge and I think the number has nothing to do with the present and future of Open Source as a method of developing software or as a classification of software.
First off, some projects die before they leave the gates. There are dozens of projects on SourceForge where the mission statement is something like "We're out to build the most technologically advanced 3d rendering system EVAH!" Many of these projects will go no where because they're not realistic. I would posit the people who created these projects are either inexperienced in the subject at hand, they're eternal dreamers, or they don't really know what it is they want to do in the first place. There's nothing wrong with that--most of the developers I know cut our teeth on things like this and learned valuable lessons as to what ideas we have that are worth tackling and putting the energy into and what need some more time to be fleshed out.
This happens all the time in non-Open-Source projects. Watching any software company's press releases over time alone shows that some projects are good ideas and some don't make it past the gates.
Second, some projects lack the infrastructure and management to survive. Most of the projects I've worked on are dangerously close to this line. Every project that is successful has a huge amount of development and user infrastructure there to facilitate communication, testing, bug-finding, documentation, and a slew of other things that healthy projects need. Why do projects need these things? They make it possible for a diverse group of people to work on the project and communicate with each other. They make it possible for users to look at the project, figure out if it fits their needs, install, configure, use the software, and report issues back to the development team. Without this infrastructure in place, it takes a huge amount of energy to keep things running and unless people are working on the project full-time, it's likely people don't have time to pull that off. I would posit this is the biggest issues for small and medium Open Source projects. Building infrastructure and policies is hard work and it really requires the people involved to have a lot of experience in this sort of thing.
This is true of many things--wood-working projects, making movies, starting up a band, starting a restaurant... What is the success rate for bands? How many fail or disband after a few months? Does that mean that music is dead or failing?
Third, some projects are abandoned because similar projects end up being better in some way: features, management, infrastructure, momentum, marketing, etc....
This happens in non-Open-Source projects all the time. When was the last time you heard of someone using Lotus 1-2-3 or VisiCalc?
While not comprehensive, I think these three reasons cover a huge number of failed/dead/abandoned SourceForge projects.
This isn't a scientific study. My experiences with Open Source software has led me to believe it's totally ridiculous that people point at the failed projects on SourceForge and use that to base their ridiculous claims that Open Source is a fad and/or that it's a failed model and/or that it's going to die.