A while ago there was the thought that plexnet could be built on top of PSYC technology. But as long as it is not clear what needs to be built, this is a pretty vague thought.
Applications of plexnet
Here a list of example applications of plex taken from http://esp.k42.org/plexnet/
* An automatically update address book
- This in PSYC would be done with the decentralized state system, that isn't available yet, but I'd like to dig into it.
* A spam and virus-free discussion system
- See SPAM on the PSYC approach for that. According to roadmap it just needs some bug fixes and tuning trust and we have that from a back-end point of view. Then you still need to plug whatever front-end you want on top of it. I guess the time is right to start doing that front-end, if you care.
* Discussions with summaries and ratings to reduce information overload by allowing users to read only the key parts
- <jeffarch> trust+trading facilitates "who is going to do the work of summarizing", summarizing is tied to "indexes"
* An accounts system with ratings and reviews to allow people to trade with those they trust on the basis of quality
- Similar to trust maths which should be employed on a higher abstraction level to all kinds of things. I wouldn't mind to extend the protocol by more attributes if they make sense.
* Trading systems with built-in local currencies and reputation measures to reward public service
- Sweet. sounds like a nice application.
* A flexible file-sharing system
- Yep. File sharing in social networks etc. We're fixing our client protocol, but essentially we could be going there real soon back-end-wise, if you join in. PsycZilla, psyconaut and even psycion are getting close.
* A 'meta-aggregator' to pull together information from lots of systems such as chat, email, photo sharing, etc.
- We poll by HTTP if unavoidable, but we prefer to parse SMTP. Even better.. gateways and newsfeeds. Don't forget to multicast information! In other words, yes we are doing this.
* 'indexes' providing generalised associative metadata, so that everything is searchable and categorisable
- That's an interesting extension we have not been planning for, so far. Sure it is useful for searching in a decentralized file sharing network. When is it correct to search (aka poll) for something when you could have friendcast it's existence in advance? Where is the fine line where multicasting events is less efficient than discovery?
24weeks.com was a hackathon run by the espians some years ago. The domain is gone so it's unclear what the results of the event were.
Update 2011: http://inamidst.com/stuff/esp/ampify is the new big espian vision although I can't say what exactly happened to the ones that went before. This time it reads like application messages that could be done on top of a messaging system, probably the plexnet which partly exists on github. There's a text template strategy similar to psyctext.
In the end the most important thing about ESP is the people.