I tend to think of politics as resource arbitration. Say, you have limited resources (as we all do) and there are too many contenders who want to settle this without resorting to any sort of violence, then you need pretty much politics.
Technically, a kernel scheduling algorithm or packet forwarding algorithm are politics too (hence they're called "policy"). The problem is that the real world is so messy and complex and has too many variables unlike CS stuff that we can hardly reach any sensible solution in a timely way. In theory we can tackle on politics in a somewhat objective manner but it's typically waay harder than any software project.
See the book: "Policy Paradox: The Art of Political Decision Making" by Deborah Stone for an exploration of this issue of resource arbitration and politics.
One simple example she uses is how do you divide up birthday cake? Equal pieces? What if come people arrive late to the party after you start cutting the cake and giving out pieces? What if some people are full but others have not had dinner yet? What if some people don't like the pieces with chocolate icing?
While we can be "objective" within a set of established priorities, we can't be objective about designing a set of priorities. Albert Einstein wrote on this in "Science and Religion" about how science can tell you about what is and how it is all connected, but it can't tell you what should be.
One other point Stone she makes is that when people are on the outside of an organization they want transparency of the decision making process -- while then the same exact people move to the inside of an organization they suddenly have many reasons why they want decision should be opaque. "Yes, Minister" has a great comedy episode on that called "Open Government".
Technically, a kernel scheduling algorithm or packet forwarding algorithm are politics too (hence they're called "policy"). The problem is that the real world is so messy and complex and has too many variables unlike CS stuff that we can hardly reach any sensible solution in a timely way. In theory we can tackle on politics in a somewhat objective manner but it's typically waay harder than any software project.