Thanks for stopping by and commenting – excellent points on both counts! I did struggle a bit with the OSX example, having realized it was a hybrid – however based on your comment your knowledge is much deeper than mine in this area. In the end I went with the rationale that one must purchase it to use it as you had mentioned. This does imply a closed nature, although less than other operating systems. I do agree that there is a great benefit to the hybrid nature, especially when stacked against the competition.

Concerning gatekeepers I couldn’t agree more. Gatekeepers are essential to maintaining quality and directing the project by receiving and screening the results. One cannot simply announce the project and step back – monitoring and direction is required. The overall benefit here is that many eyes can solve a problem faster than just one or a select few – especially if the few sets of eyes are from the same organization. Diversity is beneficial in this instance. Here at this blog we have had problems solved within hours by the crowd that might have otherwise taken days or weeks to solve on our own.

The best method of maintaining quality over time is to break the task into small, discrete components. This allows for small chunks to be completed quickly and also allows for ease of monitoring of quality. Of course this pushes complexity to integration – and you know as well as I that integration is no trivial task and has its own set of headaches. If managed effectively I believe this could be a powerful tool, but would require a strong integration, management, and quality team. It will be interesting to see where this goes over time.

Thanks again and I hope to hear more from you!

I think one problem with crowd sourcing is that the output after a length of time tends toward moderate. You can’t design by commity for ever. A prime example of this is the Linux kernel. Yes, there are hundreds (probably thousands) of guys that hack on the Linux kernel, but there is only one gate keeper. If Torvalds doesn’t want something in the kernel it isn’t going to get in there. This keeps the kernel moving forward so that it doesn’t tend toward mediocrity. In fact, there are only a handful of developers world wide that can submit changes to the kernel directly (all have been blessed by Torvalds and recieve direction from him as well).

The same goes for the NASA program. They had several people pouring over the work of the crowd. You won’t get that high quality product without very knowledgable gate keepers behind the scenes.

–R

Mac OS X is not a really good example of closed source. It is actually a hybrid. The code for the interface (Aqua) is closed, but the kernel and other parts of the OS are open source.

http://developer.apple.com/opensource/index.html

The Darwin Kernel is based on Mach 3.0 and the UNIX layer is FreeBSD 5.

The rendering engine in Safari is also open (Webkit).

Some people would argue that it is essentially closed as you can’t run OS X without buying it, but I think that it benefits greatly from its hybrid nature.

–R