- Why open source app code?
If you put out source, you can get people to come and examine it / to give people some feedback or extend to something good to use.
Apple's Quick silver had bad code. The person who wrote/ managed Quick silver joined google and put out source code and said it's bad code. Users determined that performance is bad anyway vs developers announcing this. Quick silver failed even before its code was open sourced.
But an individual demonstrating bad open source is different from a Corporation / Government saying it's bad and inviting criticism. Usually happens because they may have too many unnecessary conversations going on when it's open source, which may hinder development.
- The government's understanding of software
EVM argument - Government does not understand software infrastructure because they want to think of it through their own structure. Competition will make services cheaper. Open source can't help the make the service cheaper in cases like Airlines, Railways. E-rail hasn't replaced the official system. Where there is a barrier to entry, open source projects can't enter (IRCTC)
- Ubuntu success
Successful open source works if people can build something over it.
Linux Kernel contribution is not fragmented. - Software is moving towards consolidations.
Canonical has the trademark of "Ubuntu" - Ubuntu 1 was a proprietary file sharing app like drop box. Ubuntu is sponsored by Canonical - pays for Ubuntu releases. Ubuntu has a "very well managed" community also because money gets pumped into Ubuntu. The commercial angle exists and is driving it forward.
- Extending the Ubuntu Dependencies experiment
It helps to understand the influence of packages, who is contributors, which organization are they affiliated with.
We can then determine: What is the amount of money pushed in? What is the size of the community that he has gathered?
- School Project would be interesting
School project should be open source + sustainable. We can understand bootstrapping on different layers of infrastructure and what set of users Different layers have. Clone is not infrastructure - not done well. Zope3 is infrastructure - customizable, reliable, reusable. Pantoto?
Infrastructure line is very thin...is very expensive. There is lesser money lower down on the infrastructure line and so you can't afford costs of customizing anything on top of it.