<br><div><span class="gmail_quote">On 3/4/06, <b class="gmail_sendername">spike</b> <<a href="mailto:spike66@comcast.net">spike66@comcast.net</a>> wrote:</span><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Google is your friend.<br></blockquote></div><br>Sometimes... It depends on how much one knows and how narrow the topic is. If you have precise keywords I think Wikipedia is a better place to start. If you hand "politics" to Google it gets you 900 million references. Even "artificial intelligence" gets you 90 million. In Wikipedia you get 1 entry for each or in some cases a page for disambiguation. Usually the entries have a set of references pointing you to the rest of the related "body of knowledge". With Google you are going to have to work to get those. Where Google may excel is when you have something which is at the intersection between one or more areas. For example (politics "artificial intelligence") gets you ~3.5 million pages (cutting things down by 1-2 orders of magnitude). That is still a lot to wade through but its a lot less than either topic alone. I also doubt that Wikipedia has an entry about the intersection between those two subjects.
<br><br>Google Scholar solves some of this if you are looking for the cutting edge research but this process is going to be problematic until you can request information based on your level of knowledge in an area (e.g. elementry, junior, high school, undergraduate, graduate, top 5 people in the world, etc.).
<br><br>Where it will become easier is when one can have topic-map displays of collections of information based on the concepts of "local attractors". This is similar to the "related articles" selection in PubMed but I think they had to use a supercomputer to do the groupings (on a database which only has something like 10 million short abstracts). I think I have seen it on one or more news sites as well (CNET perhaps?) but I think they confined it to their own news articles which is a small dataset. I don't think even Google has the computer power to map the document family intersections for all of the documents on the web. If anyone knows of sites which do a particularly good job on this, particularly if they display the information as a visual graph, please let me know.
<br><br>R.<br><br>