The best way to answer this question comes in two parts, three if I include Microsoft. The first part is to examine what happened with XFree and Xorg a while back. XFree changed its license, people revolted and focused more effort on Xorg. Licensing can change at any level of development to fit the needs of the community but when there's a point where the project has to fork, it will fork. How this can relate to cost will be explained later.

The second example of this is compiz, beryl and compiz-fusion. It has been forked and consolidated over the life of its development. It has done this in part to provide equivalent functionality of other OSes, if not more flexible, to the Linux community. Innovation spring boarded because of how forking and consolidation worked together. I'm sure people would be hesitant to agree with this especially if they were directly involved with the development. When you're in the middle of a battle, it's hard to see straight. The value of this was flexibility in order to develop a stable package in a short amount of time and get it so it can be built upon through a community of plugin developers. While most projects never see the light of day, this one did.

I'll include one aspect of Microsoft. If you haven't read the news for the past 10-15 years you might've missed how MS treat ideas and how they assume it into their own system. A lot of the ideas and features MS provides today come from various communities that provide a means to capitalize on these ideas and then centralize the profit without concern to the communities that support it. The cost is recouped on businesses and their ability to achieve better effectiveness because of speedy implementation. MS hasn't been as speedy as I would like them to be for quite a long while. Their development cycle has slowed a bit for my own personal tastes. In a way, they are at the will of the developers willing to provide innovative code into the fray. The developers are very leery of their own patents and how other people can profit from them. While there has been a lot of change, sometimes things stay the same the more they change.

What does this mean in the long of it for OSS? Well, the cost to distribute is a problem because development relies on ISPs. Mirroring of repos can help to distribute taxation to servers. P2P can distribute the load to the end user. Developers can volunteer depending on the availability to perform. In a weird way, their own need to express themselves through code inspired the need to find ways to divert cost.

On the other side of the coin, various personal reasons can keep people from being more active. The general economy is a big part of the equation. In order for innovation to evolve at a reliable speed, a certain level of integrity of the community has to exist. Commercial involvement would be welcome as long as the flexibility to perform remains untouched. I don't speak for the community at all when I say any of this, just so you know.

Syphoning ideas from end users can only go so far. Maintaining the integrity of the dreamer can be ridiculous at times, but can provide a means to surprise every once in a while. I'm sure there are people more than willing to do anything if they knew what that thing was. All in all it is all chance unless you find a man with a plan. I think if you find that. You find your answer. And as a reminder, this is all my own humble opinion.

I really hate bashing Microsoft at this point. I'm not going to. I will say this and leave it at that. Security within Windows can be problematic if you don't properly setup your permissions. Hacking the registry helps but you either have to rely on Windows update or a third party for any real fix. You really need someone who knows what they are doing though to do it effectively. Antivirus and firewalls don't cut it all the time when you have poor policy practices. That's true for any system.

I hate suggesting this but the US government needs to implement Linux or something similar in all their security critical systems. Linux is great for setting up local repositories for local networks.