Searching Upstream You've just found a great target web page.  You've even followed the links "downstream" from that web page to other web pages.  Now do the most important search. Find the web pages that point towards the target web page (= searching "upstream")

Look at these three scenarios and see the value in these different approaches to searching.

1. You discover a valuable web page called target.html.  Most people simply explore the hyperlinks contained within target.html.  Those links take you to places suggested by the author of target.html.  In short, all you are able to discover, are the pages which are located "downstream" of target.html
2. You discover a valuable web page called target.html.  In order to help judge the importance, value, popularity of the page, it would be nice to know "How many other web pages contain hyperlinks pointing towards target.html".  This shows how many web authors know about the target site, and felt that the site was "good enough" to deserve a hyperlink from their own web site.   If a web page looks fairly anonymous, you might be able to infer something about the source of the web page based on the who else links toward the web page.  Fortunately, some search engines (like search.yahoo.com and AltaVista) will allow you to search for all web pages which contain a hyperlink to a specific URL.  This is what I call "searching upstream" of a web page.  Here are two example searches:
           link:http://navigators.com/isp.html <-- Shows who links specifically towards my ISP Page.
           link:http://www.whitehouse.gov site:ru <-- shows pages which link to whitehouse.gov which also happen to be have a domain name ending in .jp.
   Note: some search engines can rank their search results based on the "popularity" of a web paged derived from how many links point toward that web page.
3. You discover two valuable web pages called target.html and other_target.html.  If you can find any web pages that link to both of these target pages, then you may discover a great "directory" such as "joe's list of targets on the Internet"  for the subject covered by these target pages. For example;  Suppose you want to find a list of news sources for the country of Colombia.   There is a short list of Colombian newspapers listed in Yahoo.  First I  check how many web pages link towards Elmundo and then how many link towards Elheraldo.  Once I am satisifed that these two newspapers websites are reasonably popular, I can then search for web pages that link to both newspapers.  The links to accomplish this search are provided here: Colombian Newspapers: Yahoo's List,
   - search.yahoo.com = link:http://www.elheraldo.com.co ,   link:http://www.elmundo.com , links to both newspapers.

Here are additional search examples:

In the diagram below, you will see how  your "Searching upstream" results will vary depending on which pair of target sites you use:

You may want to also try ALEXA Traffic Rankings, for similar information based on the traffic patterns of millions of users of the Alexa Toolbar.

Searching upstream is not only a good idea, it is actually necessary in order to even reach almost half of all web pages known to Alta Vista

Detailed research paper on this: http://www9.org/w9cdrom/160/160.html

Note for my Alumni: Your referrals are always appreciated