How much is enough?

The post by Academic Jungle coincided by a few occasions in my lab that inspired the following question:

How much contribution is enough contribution to have your name in the paper as one of the co-authors?

Two scenario's arose this week:

1) I submitted a paper in one of the conferences. In my field, preliminary results are generally published in conferences. I included a name of one of my colleague as a co-author; the reason being his very useful input in producing whatever we were trying to achieve. Although he was not involved in the actual 'work' that we did, his input was necessary for the system to run.

2) On the other hand, I was on the giving end; i.e. one of my colleagues had a problem that I had particular expertise in . Mind you that these guys were 'discussing' the problem for two days or more. When I was asked the particular set of questions, I added my two cents making them do what they have been trying to do for the past two days. Point to note here is that my name was not included in the co-authors despite the fact that my input was imperative for their system to work. My other colleague who was involved in the same work, but didn't had any 'intellectual' input. The only input that he had was the routine menial work that every one knows and do on routine basis(and of course is necessary for publishing the paper).

Question is how much is enough to include someone in one of your co-authors? and if the input from one of the colleague is intellectual in nature, and from another you get something like routines work that is also important for the paper; how do one quantify the relative positions of the co-authors in the paper?