Twitter Data

I like the idea of Twitter data but have concerns about how obtrusive it is. Especially considering the examples with name spaces and such, it is very much a computer language rather than a human language which is more in twitter's nature.

When I saw twitterdata.org and read about the concept, I was like, yes, great idea. But when I read about the notion of tuples and namespaces and encoding of things I was a bit disappointed. In my view something like twitter data would be great to specify RDF-like facts, e.g. I like comcast, I hate madonna. If you would write a tool that analyzes such information I can imagine it's difficult to parse and I can imagine that you need some kind of light-weight encoding to extract the subject, the verb/property and the object. I can imagine putting a $ in front of verbs, for instance. I $like comcast, @zef $livesIn Holland etc.

Your twitterdata proposal seems to take another, more complex, computer scienc-y approach. The messages written in this format only have the purpose to encode structured data and are not so much human-readable messages anymore, which is a shame.

Any thoughts?

Hi Zeff,

I $liveIn SanFrancisco and $like #zembly

is exactly the kind of things we thought people would be able to read and write. The usage of two level namespaces e.g. $foaf>knows is optional. It's d left to the bottom process to figure out what works and what not. We only provide the idea (and a parser) :-)

In your reference to RDF, notice that the above is similar to RDF N3 notation; just a bit more suited for embedding in natural text.

Jiri