I think part of the reason why programming (and open source especially) allows for very young people to participate easily is simple: bugs always exist. Open source projects need a lot people -- not just A-grade programmers or developers or computer scientists who've had decades of experience. They need documentation, bug reports, patches, active forum/mailing list members, and community organisers to be really successful. It's easy to get started just by hanging out on forums, eventually noticing a bug, reporting the bug, learning more about how the software works by understanding the bug, possibly trying to fix the bug, etc. Over time, it isn't too hard to start making very significant code or documentation contributions to a project.
In contrast, bug-testing Newton's Laws is a waste of time, and bug-testing General Relativity requires a very large budget. Attempting to make a meaningful contribution to physics is very difficult. The same can be said for any pure science.
By definition, any pure science (natural or social) doesn't evolve quickly enough to allow for quick-but-meaningful contributions because pure science is about establishing a base of knowledge that builds directly on previous knowledge. Without an understanding of the "previous knowledge," any newcomer would be lost and wondering what's happening.
In contrast, applied science is often an orchestrated effort that involves experts in different fields and contributions from different fields (e.g. your software is as valuable as its support, and your interface ideas are as good as the person implementing them). Prior knowledge isn't necessarily as valuable as the ability to integrate ideas from different places or collaborate with others.
Would you agree? I ask because part of me feels like I might be missing something.