We all have a vague, intuitive sense of what the appropriate degree of publicness and privateness is in a given situation. Although vague, our sense of privateness is incredibly fine-grained and what is perfectly normal in one context can be a massive violation in another, almost identical context. While some of us geeks might think of "public" and "private" in a binary, cryptographic sort of way, most people have a much more nuanced approach.
Consider for example the difference in privateness between two people sitting on a quiet park bench and two people sitting on a bus seat. They are both obviously public places, but there is a subtle but significant difference in our expectations of privacy. In one context, eavesdropping is perfectly normal (within certain bounds), in another it is quite sinister.
On the internet, these intuitions are frequently confounded. Someone who assumed that their facebook feed was fairly private discovers youropenbook. User 927 assumed that his search terms were just noise in the crowd until AOL published them all. Countless iPhone users didn't know that their photos contained their exact geographic location until 4chan had a go at ruining their life.
The problem isn't the level of privateness or publicness of a particular service, it's not even particularly about leaks or breaches; it's about people being completely unable to judge the level of privateness of anything digital. I expect most people simply would not use a 100% public medium for their private communications, so the "if it's on the internet, it's public" mantra isn't a great deal of use. I think this is one of the reasons why Twitter has been so successful - it provides a platform that implicitly communicates its privacy or lack thereof. Twitter's simplicity makes it easy to understand and integrate into your existing model of social appropriateness.
I think we come back to a very old and very simple principle of interface design - don't surprise the user. We should be trying wherever possible to design systems that are as private in practice as they would seem to be intuitively. Users don't read much of anything, so we need to think about other ways of communicating "publicness" and "privateness" in our software. We need to recognise that designing social spaces is not a primarily technical problem. Small differences in architecture, interface and even general ambience can create enormous differences in how a platform is used. This stuff is really hard and really easy to completely cock up and I think we need to think much harder about it. Nod to PG here - HN is IMO a great example of a subtly, intelligently designed social platform.
Consider for example the difference in privateness between two people sitting on a quiet park bench and two people sitting on a bus seat. They are both obviously public places, but there is a subtle but significant difference in our expectations of privacy. In one context, eavesdropping is perfectly normal (within certain bounds), in another it is quite sinister.
On the internet, these intuitions are frequently confounded. Someone who assumed that their facebook feed was fairly private discovers youropenbook. User 927 assumed that his search terms were just noise in the crowd until AOL published them all. Countless iPhone users didn't know that their photos contained their exact geographic location until 4chan had a go at ruining their life.
The problem isn't the level of privateness or publicness of a particular service, it's not even particularly about leaks or breaches; it's about people being completely unable to judge the level of privateness of anything digital. I expect most people simply would not use a 100% public medium for their private communications, so the "if it's on the internet, it's public" mantra isn't a great deal of use. I think this is one of the reasons why Twitter has been so successful - it provides a platform that implicitly communicates its privacy or lack thereof. Twitter's simplicity makes it easy to understand and integrate into your existing model of social appropriateness.
I think we come back to a very old and very simple principle of interface design - don't surprise the user. We should be trying wherever possible to design systems that are as private in practice as they would seem to be intuitively. Users don't read much of anything, so we need to think about other ways of communicating "publicness" and "privateness" in our software. We need to recognise that designing social spaces is not a primarily technical problem. Small differences in architecture, interface and even general ambience can create enormous differences in how a platform is used. This stuff is really hard and really easy to completely cock up and I think we need to think much harder about it. Nod to PG here - HN is IMO a great example of a subtly, intelligently designed social platform.