Wednesday, April 8, 2009

Finding Interesting Information on the Web: It's Not Just for Search Engines Anymore

Today, I'm going to take a break from Mindmapping to take a look at an emerging area of technology that I find exciting and believe to be tremendously promising down the road. The related technologies that I'm talking about are related to search engines, but have evolved enough in various directions that they really should not be considered search engines in any traditional sense.
This train of thought started when I read an article in ReadWriteWeb about Ambiently. Ambiently, which is currently in beta, calls itself a discovery engine, as distinguished from a search engine. What Ambiently does is to find web pages that are related to a given page. It is implemented as a bookmarklet. In order to use Ambiently, simply go to their Web site, drag the bookmarklet to your bookmark bar, visit a Web site of interest, and click the bookmarklet. You'll be presented with links to pages that Ambiently believes are related to your original page. At the moment, what you'll find a mixture of expected and unexpected links. The unexpected links are unexpected for one of two reasons:

  • The link doesn't lead you to a page that is related to the original page in any meaningful way, or
  • The link leads you to a page that contains information related to the original page, but in a way that you probably would not have found using obvious search terms with a traditional search engine.
An example of the first case, where a link is not meaningfully related to the original page came when I clicked the bookmarklet while viewing, which is the beta version of FriendFeed. The Ambient page had lots of links to pages about "beta." However, I spoke with the developers, and they are working on it. An example of the more useful case of unexpected results came when I clicked the bookmarklet from the ReadWriteWeb article about Ambiently. There were obviously a number of links to pages that discuss Ambiently, since the original page was about Ambiently. However, there were other links that I found interesting because they discussed other, similar types of application, with the only connection that I could see was that the new page referred to "bookmarklet."
I'd like to give an example of a person that would find this very useful, me. I'm a scientist. Many times I am doing research for a whitepaper or journal article. It is not always clear what search term would provide me with the most fruitful direction to go. Instead of trying a large number of search terms looking for the best one, I just have to go to a page that looks interesting, and let Ambiently do a more general search, and I can see a number of results, hopefully finding one or two that look worthwhile. At some point, I will probably end up feeding a term to a search engine, either at the beginning to find the initial page of interest, later in the process after narrowing down my research direction, or most likely both.

In my view, this idea of looking in a lot of different directions without my having to try a large number of search terms individually, provides a discriminator between a search engine and a discovery engine. Each direction that Ambiently (implicitly) looks does not provide me with a long list of links. It doesn't need to. It is enough that it gives me a peek in each direction, so that I may decide which ones to explore. Once I visit a page linked to by Ambiently, I have (again, implicitly) chosen a direction to explore. From there, I may find that the new page has links to others with useful information, or I could have a search term to feed into a search engine, or I could even let Ambiently discover some new directions for me from this new starting point.

From what I've seen so far, the reviews have been somewhat lukewarm, but I think that there is something much bigger going on here. I am really excited about this technology, partly because of what it can do for me today, but mostly because of the tremendous potential I see for these alternative approaches for finding information on the web.

There are a few other technologies that I do not see as search engines in any traditional sense that I am going to mention, but not discuss in any detail (perhaps some other day).
Primal Fusion, which is currently in alpha, has a concept of "thought networking." Here the process is more iterative, and can build over multiple sessions and multiple searches. You can start with a search topic of a few words. This brings up a tag cloud of "thoughts." These thoughts can be "remembered" or clicked upon to search and retrieve content. The content can be remembered, as well. If your current research is done, you can logoff and go about your life. Then, when you login to Primal Fusion again, your remembered thoughts are still there. You go through the same process as before, adding further thoughts to your memories. These remembered thoughts show how you think about the web. Over time, the thoughts are re-organized, and refined. New thoughts can be generated, providing new insights into your thought process. Once you have done all of your research for a topic, you can generate a Web site about the topic, create a document, or create an RSS feed.

evri, which is in beta, starts out with a more traditional search, but then allows you to navigate through a network of topics/concepts. Infolust, which is in alpha, provides links related to a page, similar to what Ambiently does. However, it appears that Infolust currently only links to Wikipedia articles.

I believe that this is only the beginning of the changes that will be taking place over the next few years regarding the manner in which we explore the web. Feeding a search term into a search engine is powerful thing. I would get very little done today without it. However, as we move forward, it will just be one of many ways in which we find what we want or need on the web

No comments:

Post a Comment