So you almost had me there. First of all your points are all valid. Where something felt wierd to me was the edges. What is the exact value for customers in this edgeset being maintained and worse harvested, I mean processed? Today we have edges outside the context of a social network - my contacts in email, phone book etc. And those "edges" (not the target node) belong to - you guessed it - me. Nobody should harvest it without consent and/or maliciously. (There is the whole argument about internet ceasing to exist without ads and nobody would pay yada yada which I felt was too reductionist). If somebody needs to harvest it, get consent and let user decide how, where, when why etc.
> And those "edges" (not the target node) belong to - you guessed it - me.
For a contact in an email adress book, that makes sense. But for a "friend" relationship in Facebook, which side owns that edge? Or how about a message sent from someone in the EU to someone in the US, who owns that, the sender or the recipient? And if it is just one, does that mean that different messages for the same conversation have to be stored in different regions?
In this case the problem can be solved with 2 edges :) I am your friend and your are mine. Keep an edge on each side. Heck I could be your friend you may not chose to be my friend and that is fine. This gets even more fun as now both parties have to consent to only share "their" friend status with FB. Americans are forced to share their friendships, Europeans are not. Again total value for users no?
Now is this technically optimised (for the company) - no and irrelevant (IMO) in the context of how much control/power a user has. You could extend this to messages too. What messages I sent, what messages I received. I didnt send it - I dont own it. What about shared documents you say? Here users are explicitly sharing with other users for collaboration (the contents of said documents totally are of no business to the company).
See providers are providing a service(?). If the services needs to harvest data I still question who is benefiting from that harvesting? If the user is not actually seeing value (apart from subsidizing the cost of the internet) are we then not just using technical/UX complexities to justify a low-value (to the user) solution?
> For a contact in an email adress book, that makes sense. But for a "friend" relationship in Facebook, which side owns that edge?
I don't see where there's any ambiguity on this issue. Each individual has the right to not be subjected to spying and monitoring, which includes collecting personal and private information. A social graph is not a data dump where you are a mere drop in the ocean. A social graph is an ocean of personal and private data collected from you. Therefore, it's quite obvious that individuals have the right to not have all this ocean of personal and private data collected on them, specially without their explicit and informed consent, and they should have the right to force anyone to delete this info, both all or subsets, automatically and reliable and verifiably.
Just because I don't mind hearing what my aunt has to say about what she baked or who she chatted with, that does not grant you the right to get my credit score or where I went to highschool with or who I met years ago or where I lived, just because third parties and other edge nodes in a social graph posted that information and data that enabled you to piece it together. What is there to be discussed?
So in this context is your con really a con?