On the simplest level, iCites are lists where each item on the list links to some information, plus has some attributes assoicated with the link. These links can link to other iCites, or to any addressable information (e.g., URLs on the Internet).

A blogroll (or blog roll or blog rolodex) is a list of blogs. Usually, one keeps, on their own blog, a blogroll of other blogs they read. (Note: the iCite net development blog is so new, that it does not yet have a blog roll--it will have one in the future.) As an example of a blogroll, take a look at the right side of Sam Ruby's blog, in the section called "blogroll".

Since iCites are lists, and blogrolls are lists, it should be pretty obvious that an iCite can be a blogroll. But, what is interesting about iCites as blogrolls is that iCites are query-able, iCite links can have attributes like a rating or a category name, iCite links to other iCites are automatically tracked in both directions, and iCites can aggregate the information that lives on the ends of its links. So, let's look at some ways this can add up:

With iCite link attributes, a blogroll can have each link categorized and rated. Categories can be express via either hierarchical or faceted classification schemes. Ratings can be simply the iCite default ratings (from -1 to 5, with -1 meaning "not recommended", 0 meaning "not rated" and 1 through 5 being degrees of positive recommendation) and/or other rating schemes includes ones you can create yourself. If you want to rate blogs like "airbag", "paranoid android", or "subterranean homesick alien"--go for it. (Note: iCites can and usually do publish their attributes so others can use them. So, if you create a rating scheme like that, other people can adopt it too.)

With the iCite's query and aggregation facilities, you can show your blogroll in multiple views--and your queries can span multiple iCites. So, you can show a blogroll that not only lists each blog, but also shows how many sites on your blogroll link to that blog, say like this (note: links don't do anything):

• A blog everyone reads (12 readers)
• A blog only a couple read (2 readers)
• etc.

Additionally, aggregation and queries can use the attributes on each link on each iCite. So, one example would be aggregating a score for each link on your blogroll based on the weighted total of your own rating of the blog added to the weighted ratings of each other blog's (on your blogroll) rating of the blog. This might make a little more sense with a little math illustration:

My ratings of blogs on my blogroll:
 BlogA = 4
 BlogB = 3
 BlogC = 2
 BlogD = 4
 BlogE = 0

BlogA's ratings of blogs on BlogA's blogroll that are also on my blogroll:
 BlogC = 3
 BlogD = 5

BlogB's ratings of blogs on BlogB's blogroll that are also on my blogroll:
 BlogC = 1
 BlogD = -1

A aggregate score of BlogC and BlogD between my blog (which I rate 5), BlogA and BlogB:
 BlogC = 24 = 5*2 + 4*3 + 2*1
 BlogD = 38 = 5*4 + 4*5 + 2*(-1)

so this is (for BlogD), blog score =
my rating of my blog X my rating of BlogD +
my rating of Blog A X Blog A's rating of BlogD +
my rating of Blog B X Blog B's rating of BlogD +

This kind of thing might be useful for showing your blogroll in an order that suggests "listed from more interesting to less interesting".

With the ability to aggregate actual content that each link points to, these types of queries and aggregation could be used to create "documents" that represent a snapshot of what content within a "blog community" (by which I mean some group of blogs whose creators all read each other) is being most noticed at any particular time. One example I can think of where I would want to use this facility would be when I work as an educator and I want to keep an eye out for the person whose voice is being LEAST noticed or appreciated by other students.

Also, while it is outside the scope of blogrolls, I will mention that any iCite can link to anything that is addressable. So, querying and aggregation can extend out of blogrolls and into other links. For example, an iCite could display "blogs on my blogroll that have entries from the last month referring to books on amazon.com that I added to my iCite booklist in the past week".

Finally, I should note that I only recently came across the very brilliant Technorati site that spiders blogs and provides an interface to neat ways of querying and aggregating blog information, somewhat along the lines I have suggested for iCite. Additionally, with Google creating a search engine for blogs (and now being the owner of Blogger), it is likely that the Google API might also ultimately support similar services.

iCites take on themselves the responsibility of tracking their relationships and of providing a service interface to query them. iCites can therefore more immediately, more discreetly, more richly, and in a totally decentralized manner express their relationships when compared with existing website links as seen by search engine spiders.

But, the emergence of iCites should provide Technorati and Google with excellent and well packaged information that they will be able to use to enhance their services. So, there should be very natural compliements between them and the iCite net.

Note: future versions of this example will include live iCite examples, tips and tricks, how to's.