This is speculative. It is about a totally Transparent Consumerium.
- Server logs would not only be visible to all, but, tools to make sense of them would be provided, and server statistics that reveal interest in pages, and relationship between browsing and editing, must also be provided. The Holy Grail of transparency is to see the full Markov chain of events from any angle. This is what net advertising servers are good at tracking, and good money is paid for it. Even when applied to a nonprofit project like the Wikipedia, it would be valuable to see which articles were leading most frequently to reading of other articles. See essential projects - which would benefit from authors/editors who actually knew the readers browsing path?
- When point of purchase queries are being handled by a network of servers, these must compile their "hits" into a single server log that will tell us somehow what influence the system is having on buying. That too should be public, so we can leverage the actual impact with greater potential impact. If some company is losing sales due to us, and can see that, they can respond much more rapidly to false information. If we provide that data free to anyone, no liability for publishing it can accrue - it's their fault for not checking.
- All watchlists and perhaps internal social network and contact network listings would be public - anyone could inspect interconnections and interaction and interests - call these the three ins. Of course one can always use external networks Consumerium can't see. Unless it retains opaque control over for instance the ability to contact someone else via the system - requiring anonymity or constantly changing userids or something. However this could be defeated by people simply swapping some more stable userid via the system. A dating service is a good prototype of how this happens - in general control can only be kept over rendezvous for those who do not yet trust each other.
- The hardware requirements and hardware bids to provide them would rely on modular hardware specifications that multiple suppliers could meet, but might rely on single- or double-source parts underneath (say for a server box which has high reliability requirements). At the edges of the network, where many devices are operating, modularity and interchangeability matters, so a truly Transparent Consumerium would work with wireless standards or USB only. In theory a laptop server could be carried into the retail outlet and provide the results directly from a USB barcode reader with no wireless or static data involved. That is the easiest case to work with, conceptually. The only aspect of the network that must be hidden is the IP cloud and static to dynamic server log report that updates the live servers with the content wiki and opinion wiki on them, with the reports from the field.
- Consumerium Governance would not let any party rule the policies, and make it easy for another entity of equivalent or stricter values to fork off and compete. None of this would have any impact on the technical architecture above. This might let groups like Amnesty International or Green Parties or Simultaneous Policy Organization or Greenpeace run a part of the signal infrastructure. Since the trust model is abstracted and not reliant on who has infrastructure owners trust, it affects nothing for an entity to both run the servers and provide some input to the wiki.
Separating trust decisions from technical decisions is one advantage of this transparency exercise. So is having clearer requirements for mediawiki, which may or may not still be in use by the time the live system starts up.