Advanced Search Strategies

 Search On Kierkegaard

We will use HotBot located at www.hotbot.com for the following searches. Both searches have the dropdown boxes "Look For" set to "Boolean", "Date" set to "Anytime" and "Language" set to "English".

Let's search on Soren Kierkegaard, a 19^th century, Danish, Existentialist philosopher and see what he has to say about the philosophical topic of angst. Then we'll examine the search results and see what happens when we fine-tune the search:

Enter "Kierkegaard" (without the quotes). 21,400 matches appear and most are relevant. Too many hits here so let's reduce the count.

Now enter "Soren Kierkegaard" (with the quotes). 4,200 mostly relevant matches appear. Still too many results and too many books for sale keep coming up in the results. We want online texts and information on Kierkegaard and his philosophy. Since he was a philosopher let's add a search term.

Now enter ("Soren Kierkegaard") AND philosopher. The results are pared down to 1,600 matches.

Since he was an Existentialist philosopher lets add another term:

(("Soren Kierkegaard") AND philosopher) AND Existentialism. This narrows the search to less than 500 matches. Almost all are relevant.

Now let's see what Kierkegaard had to say about angst. Let's add another term.

((("Soren Kierkegaard") AND philosopher) AND Existentialism) AND angst

This produces 30 matches on our topic.

 What is the point? Simple. Start with the most basic terms and whittle away at the results. Add search terms in the order of topical relevancy one by one to see what happens. You'll get a better idea of how materials are being described and will be able to better control your search.

Of course the search example above could have been entered like this:

(("Soren Kierkegaard") AND philosopher) AND (Existentialism AND angst)

It's easier to keep adding parentheses and terms as you need them. Remember when using parentheses in a Boolean search that there must be as many left parentheses as right parentheses. In other words they must balance evenly. It gets confusing to remember where your parentheses go so starting off small and working your search up helps eliminate this confusion.

Search On Linux

Let's try a more complicated search. Say you run the Linux operating system and need information on building a more secure computer system using a firewall on Linux. Let's try to find "How-To's", tutorials and helpful information on Linux firewalls.

Enter the term Linux, just for laughs, to see how many hits result. 3,473,400 matches were found. That's a lot of material to wade through!

Now enter the search this way:

Linux AND ("computer security"). 22,600 matches. Still too many to deal with so let's refine this a little further.

 ( Linux AND ("computer security")) AND firewall. 7,500 matches. Still too many. Let's see if we can find "How-To's" or tutorials on the subject. Nothing readily obvious is listed in the search results that would be of help to us.

( Linux AND ("computer security")) AND firewall) AND (howto OR tutorial). 2,800 matches. Still too many hits and none are of obvious help.

Now lets try this:

 (( Linux AND ("computer security")) AND firewall) AND (howto OR tutorial)) AND FAQ

This yielded 2,000 hits. I'm still getting too many Microsoft related hits. So lets do this:

((( Linux AND ("computer security")) AND firewall) AND (howto OR tutorial)) AND FAQ) AND NOT Microsoft

This reduced the hits to 300. We still didn't find the information we wanted easily.

Let's change the above search strategy a little bit. Let's do it this way:

(( Linux AND ("computer security")) AND (firewall AND (tutorial AND FAQ)). 1,300 matches mostly irrelevant matches.

Let's try again!

(Linux AND ("computer security")) AND (firewall AND (tutorial AND FAQ)) AND NOT Microsoft

88 matches. None tell how to setup a Linux firewall.  

Redo the search again. Let's try this:  

("Linux firewall"). 5,400 matches that are much more relevant.

Let's try one more thing.

 ("Linux firewall") AND help. 2,800 matches. This yielded the needed information in the top 20 listings and most were relevant.

Summary

What was the point? As you can see we tried some complicated Boolean searches on Linux with little luck. The last one had the sites we needed listed on the first page and was a simpler Boolean search. Don't presume a fancy, complicated strategy will always work.

 Sometimes Boolean can be overkill and work against you. As you noticed the main terms we were interested in were "Linux" and "firewall". Once I grouped them together as a Boolean phrase a lot of garbage hits were eliminated. We needed "help related" information so "help" was the last term we needed to add.

Of course the search could be refined more but why? Once you have all the information you need listed in the search engine results on the first page there is no sense in refining the search further. Only further refine searches if you need more information.