Ego is to social networking what P2P is to file-sharing. Like in P2P networks, each user uses a program, dubbed “agent”, that discovers and communicates with other programs in the internet by following a well-defined and extensible protocol. Unlike P2P networks, agents are indexable by search engines and as such can discover each other by doing a simple web search. This eliminates the need for “seeding” nodes, directories, central servers, etc. Agents allow their owner to select the levels of access different groups can have to private information, search engines included. Agents can also run arbitrary, downloadable applications that extend their capabilities, like keeping track of their location, interfacing with their owner over IM, etc.
-
Recent items
Recent Comments
Blogroll
6 Comments
Any information on progress for this project? The new video presentation sounds like this is already available, but still no information on how to get a Agent.
We’ve been in the process of moving between servers and working on a new, lighter version of Ego. If you ‘re interested in seeing what the current version of an agent looks like, I can give you access to an existing agent in private. If you are interested in getting your own agent, I expect the website to be back up in a couple of days. Thanks.
Any updates?
Ego might be the most hidden social media solution Ive found yet, but if it can do half of what website says it can, this thing will be big.
In the video it is stated that the project is available for download.
Is this so?
Indeed, the project’s code used to be available for download. With the advent of technologies like tornado and websockets though, the original code is rather an embarassment
The project was inactive for a while but we’re now almost ready to put it back into public – with several modifications from our previous approach. The most important ones are:
- No user interface! I’ve come to realize that it’s more important to have a web-based agent that provides JSON output and a clear API, rather than contriving an HTML-based interface that uses templates. You can still use “stand-alone” HTML apps that leverage our restful interface.
- Usage of key/value pairs – we’ve abandoned the relational database approach for the simplicity and portability of a flat, key/value pair structure and, guess what, we don’t miss it at all!
- All API are centered around the user, not the application – By following a /USERNAME/APP approach for all calls we have significantly simplified the process of creating and reusing applications and it becomes easier to migrate user agents between different hosts.
- We moved from django to tornado so downloading and running the code requires no installation (tornado’s tiny code footprint is included in the source code).
more to follow soon, including the source code.
p.
Thanks for the info – look forward to the release.
The technology impinges strongly on the current PLE/VLE (+ elearning portfolios) debates in educational technology.
All the best,
Brian