INFO 202 Fall 08 Blog

GNIP is an intermediary service for interchanging data among different Web2.0 APIs without actually pulling content from the source on every transaction. By acting as both an interchange and a intermediary storage location, the service improves latency, decreases polling, and appears to be working to standardize metadata between the services. According to RWW, “It’s about scalability” and “it sounds like a great idea.”

While I’ll confess to not being fully versed in the backend magic, I can definitely see value in a system that acts as a blackbox and does transforms between non-standard implementations. This seems especially beneficial considering the speed with which new producers/consumers companies are emerging on the scene.

GNIP’s technology allows “data consumers can get complete public data streams for Twitter, Digg, Delicious, Six Apart and others without ever visiting those sites or accessing their individual APIs, subject only to the terms of service of those services. And this data can be gathered via a REST-based API or the newly launched XMPP support.”

A nice bonus is that the service is free for all non-commercial users and commercial users who are “tracking more than 10,000 people and/or rules for a certain data provider”.

Articles @ RWW and Techcrunch

As we learn about the traditional categorization systems of IO (Svenonious, Lakoff, etc) and contrast them with the tools of next generation systems like XML and tagging, my thoughts continue to hover about a personal interest of mine which is the “ethics” of journalism and (failed?) attempts at both political and journalistic neutrality. In relation to IO, I am curious if established social definitions of terms, phrases, and colloquialisms (ex. lip product on a farm animal) can be used to deconstruct what is “meant” by an author, news feed, source, interviewee, speaker, etc.

Might a database that tracked the core metaphors implied by words and phrases somehow serve as a dissecting blade or is it up to an active community to closely watch the feeds and tag at will? Can an algorithm parse the subscripts of language? In light of our reading on machine translation, tracking bias seems like an even more distant and difficult goal.

So, two interesting tools showed up in my feedreeder this week and I thought it would be interesting to hear any opinions you all might have.

1) SpinSpotter – a Firefox plug-in that…uses “professionals” to build a set of rules, an algorithm to parse articles against said rules, and a community feedback model to track and tag.

Here is a BusinessWeek article about it.

I’ve installed it, but found that since it is in Beta, there doesn’t seem to be much feedback in the system yet.

2) And… Cognition, a Natural language processing database by Cognition Technologies. They claim that it is the “the largest commercially available Semantic Map of the English language.” RWW Article

They state that their system “Understands the meaning within the context of the text it is processing”, which seems promising. If a series of “loaded” words appears in a phrase, is that a sufficient index to call an author biased?

Anyway, if anyone has an interest in this area of information work, I would certainly like to talk more about it.