> The problem is defining what constitutes as being within the scope of a legitimate operation Of course, the final authority of interpretation should lie with

> The problem is defining what constitutes as being within the scope of a legitimate operation Of course, the final authority of interpretation should lie within the public. But it should be defined very narrowly. Basically, if public cannot reason about it, it shouldn't be a secret. Specific details can be a secret. In cryptography, this problem doesn't exist. There is a distinction between algorithm and key. I don't see why this shouldn't apply to accountability of government as well. > which seems to amount to outlawing private agreements to not divulge certain information Again, such private agreements should be very very narrow, and defined by what they can do, not by what they cannot do. In particular, they shouldn't prevent talking about events that happened to you. > Constitutional free speech exclusively concerns the governments right to decide what you can say Yes, and it shouldn't. Government is just another type of social institution. Just limiting government doesn't equate to freedom. Free speech, in my view, should cover all social institutions, including private companies and religious organizations. > Your envisioned platform is essentially wikipedia It isn't. First, I am talking about primary, not secondary sources of information. Second, Wikipedia isn't very anonymous. Third, there is no automated (or collective) assessment of trustworthiness of the sources. The disagreement problem can maybe be resolved, but in a different way than on Wikipedia. My thinking was that the system would have 2 types of information, sources (primary information) and edits (reviews of information drawn from sources). Sources would be optionally anonymous, while edits would be public. You would be able to "invest" your trust into an editor (producer of edits) in the system. Editing the information would be then redistribution of the trust. This would result in calculation of how much you should trust the primary sources, based on their reliability when reporting other things. This would allow you to get a "consistent" view of reality. I am not completely sure how such a system should work, and how secure it would be possible to make it. But it's not at all like a Wikipedia.