Jakub and Greg are both right, which is why services like Github are so crucial and popular. I can easily (and publicly!) fork any project hosted on Github, make changes, and share them back, all without commit access. And anyone that’s interested in the primary project can easily see that I’ve forked the project and have changes available that may or may not be folded back into the primary project.

Github isn’t just providing ease of use: they’re creating a social network of software developers where the social currency is code .

As for git vs. SVN, our product’s proprietary with a small number of in-house developers. The dev team uses git, but the rest of the company’s moving to SVN.

With distributed version control system you have ‘private branches’, but in centralized version control systems you have very severe barrier to entry: commit access. So if you complain about ‘changes in the dark’ in DVCS, please take into account that with centralized VCS, because of lack of commit access, they wouldn’t be hard to find: they wouldn’t be visible at all! Your argument here is similar to the RIAA/MPAA argument that each copyright infringement copy is one les bought copy; this is simply not true.

The problem with bolting-on distributed features on top of centralized VCS, like SVK does with Subversion is the “leaky abstraction” problem. SVK is slow, complicated, fragile, does offer replication and off-line commits but not full distribution (where all repositories are at least technically equal). See also http://ldn.linuxfou…rule-them-all-part-1 and followups

Another interesting fact is that 70% of netbook consumers are women, many of them buy their first own computer. I’m interested in the Vivienne Tam edition I think there may be many aspects we don’t see yet, and netbooks or sub-notebooks, they are a good idea, for now. IMHO.