Information Availability(last updated: 20/4/05 10:17 UTC)


Information availability is important and currently quite lacking.


We live in the information age.  However, all the data on the web isn't as "at your fingertips" as people would have you believe.  Googling for something takes time, and the results are often not-so useful.  You have to sift through a lot of irrelevant data to get to what you want.  With a little bit of work that would ultimately pay off big-time, this can be solved.

The Problem

The information available on the Internet is not organized well and is often not easily searched.  Google made a huge leap forward in search technology, but I believe the solution lies in a different approach.  Last year (2003-2004), a Google representative came to campus and presented about Google's operations.  When I asked if Google were considering the system I outline below, I got a definite response: no.  I can think of two reasons this is the case: Google thinks the below is a bad idea or Google does not think the Internet is ready for it.  I believe that it is a good idea (lots of people agree) and I think that the proof of concept is what is needed.

A Solution

The best solution is to redesign the Internet so that it is maximally effective in answering human questions. What stores are available at this hour?  Which stores sell this product?  What do people think of the different dentists within a ten mile radius?  In order for computer algorithms to answer these questions, they need access to data such as addresses, hours, product types, and so forth.  As far as I know, a sophisticated AI would be required in order to perform the described task.  A better solution is to include metadata in web pages -- this data would describe the different information contained in web pages, allowing algorithms of reasonable complexity to process this information and provide useful analyses.

The solution just described is known as the Semantic Web.

There are two ways to solve the problem: include metadata in each and every website we wish to process, or concentrate all of the metadata in one site.  I believe that the latter approach must be done to prove that the fundamental idea is valid, as dictated by the return on investment problem.

I have written some online collaborative bookmark software called WebMark as an attempt to solve part of this problem.