Michael has taken some of the thoughts that Neil, Cameron and myself have had recently and come up with something even cooler, a research stream. The idea, described here is really interesting. He has used an RSS WordPress plugin, research activity from Twitter and feeds from Flickr (appropriately tagged), CiteULike and his blog (again only research related material). What’s cool is that those Twitter messages are generated from a script that sends Subversion log messages to Twitter (oh so cool and geeky).
This led to a follow on thought. Right now, we are at a stage where, on the web, we have platforms, aggregators and destination sites. We can consume information passively, but only access it actively. It would be very interesting to see how we can go from passive streaming to passive stream processing. In my mind that’s a Semantic Web problem, where we can find relationships in data streams, but of course, I could be completely way off base. That said, until the web becomes a platform not just for automating information distribution and aggregation, but also for autodiscovery and autointelligence, it’s not quite the comprehensive platform that it can be. That’s what an intelligent backend should be able to do; interpret all that data meaningfully with minimal human intervention.
Technorati Tags: Bioinformatics Zen, Lifestreaming, Research Streaming
4 Comments
Couldn’t agree more. This is kind of what I was trying to articulate with my post on feed tools. I think that a lot of the tools are there to generate the feeds, or clever people can write new ones with relatively little problem. We can then aggregate those together. But what is missing is the autodiscovery and semantic filtering. Also we need better ways of getting the semantics in there than we have at the moment.
A combination of BioMedExperts, GoPubMed, and citeulike, can probably put together a pretty good list of people and topics I should be following. But how you build the intelligent agent that then filters all the stuff generated I don’t know. And how you embed (and get people to actively embed) the meaning that will make the filtering possible I really don’t know.
Oh, and we have to persaude a lot more people to start generating feeds. Mustn’t forget that one
Couldn't agree more. This is kind of what I was trying to articulate with my post on feed tools. I think that a lot of the tools are there to generate the feeds, or clever people can write new ones with relatively little problem. We can then aggregate those together. But what is missing is the autodiscovery and semantic filtering. Also we need better ways of getting the semantics in there than we have at the moment.
A combination of BioMedExperts, GoPubMed, and citeulike, can probably put together a pretty good list of people and topics I should be following. But how you build the intelligent agent that then filters all the stuff generated I don't know. And how you embed (and get people to actively embed) the meaning that will make the filtering possible I really don't know.
Oh, and we have to persaude a lot more people to start generating feeds. Mustn't forget that one
Actually that’s probably the most important part. Without the collective intelligence, it’s probably not going to mean that much.
Actually that's probably the most important part. Without the collective intelligence, it's probably not going to mean that much.
2 Trackbacks
[...] Further reading Research Streaming [...]
[...] The service positions itself as a “social, stock microblogging service”, which uses Twitter as a content production platform, and then adds on to it a stock-centric structure. This immediately appealed to me, since it’s an excellent example of the data finds the data, then people find the people. I believe the way it works is around stock symbols, being able to track them via the Twitter API. What especially made me think about this were Pierre’s blog post and tweets about using Twitter for bioinformatics. This fits into some general paradigms as well, those around harnessing information streams. [...]