We turned a corner last week. We got the Blog Neighbourhood Analyser to work well (and much faster). We also did a very preliminary auto-categorisation of blogs, based on the information that’s available. We may relook at this a little later, but for now that will suffice. Hoping to do a launch of BlogStreet soon (finally).
It will have the following components:
– Blog Neighbourhood Analyser: Given a URL, it gives the related blogs based on the blogrolls. Two options: real-time if the blog URL exists, or offline as an email if the blog does not exist. The latter service would be quite useful for bloggers who want to know which other blogs they shouldbe referring to, given what their blogroll is.
– Top 100: the top 100 blogs listing
– Blog Search: on the home pages of the blogs; simple keyword search; blogs botted daily
– Auto-Categorisation: this may not be great, but am hoping it will improve over time. The objective here is to identify the blog clusters. It is very difficult to actually categorise a blog. Given a blog, it should be possible to navigate through its linkages. Like navigating the universe through galaxies, solar systems and planets. Hub Blogs (the most popular ones) are the equivalent of galaxies.
– Iterations of the BlogBot and Blogroll analysis: so that we can keep growing the number of blogs. We are currently at 1500 blogs.
The next major development is going to be the BlogPost analysis. Given a blog, we need to get its archives, and from those pages, the actual posts. When we do a search, the posts are the ones we should be pointing to, and not the actual pages.
A few other ideas:
– provide an RSS-ifier
– think about using IM for notifications
– how to apply these ideas on blog analysis within the enterprise
– how can we open source this
– providing a neighbourhood service
Lets divide the world into Bloggers (B) and Non-Bloggers are of two types: the famous ones (1% of the Bs) and the rest of us (99%). The rest of us bloggers have a URL for our blog. One of things we’d like to know is our blog neighbourhood. This gives us the cluster of blogs which become our frame of reference. We want to use the cluster to discover new items of interest, new people we want to interact, etc.
NBs are 1000x the number of Bs. NBs want to find interesting ideas/people in the world of blogs. The starting point for them into the world of blogs are (a) the Top 100 blogs (b) search, which points them to a list of blogs. NBs can then set up their own “cluster” of blogs which they like and want to track. They can do two things: (a) use this cluster as the defined search space for keywords (b) create a n RSS aggregator and a private blog, where they can save relevant posts with their comments.
In both cases, given a set of blogs, the system can also do the following:
– provide a “what’s hot” among them – links, keywords
– show newcomers in my neighbourhood
– track for new search results on pre-defined phrases / keywords
– show Amazon books which my cluster likes
The idea is to use a “trusted set of bloggers” (as Steven Johnson has called them) as information filters. This is the foundation for a knowledge sharing system, and a personal information management system.
A few questions to ponder:
– How many people would be interested in a service like this. Few people are interested in reading and fewer still in writing.
– This process calls for changes in the way people have been working.
– Will people post items? Most people still prefer to keep information to themselves.