|people in facebook or personal address book with whom one has had more than 10 interactions in the previous year||Friends||12|
|people in facebook or personal addressbook with whom one has had between one and 10 interactions in previous year||Acquaintances||40|
|People in facebook or personal addressbook with whom one has had no contact in the previous year||Interactives||~850|
|People not in facebook but either in personal address book or work addressbook with whom one has had a non-personal exchange in the previous year||Contacts||1000+|
So one can break this down to say:
Friends: real friends and family with whom one has a set of interactions - strong ties
Acquaintances: real acquaintances with whom one has a set of interactions spreading out over time - weaker ties
Interactives: people that you may have interacted with in the past and may interact with again - weakest ties
Contacts: people with whom you have no ties but have a set of interactions such as a business relationship with - no ties
This model is nicely hierarchical if a bit arbitrary. Contacts may become Interactives. Interactives may become acquaintances. Acquaintances may become interactives and so on and so on. And while salesmen always want to be your friend they're not as the interactions are non-personal.
The next stage is to try and apply this model to a larger data set ...