Searching the Internet

Talking about

• Finding Information
• Finding Files
• Finding Yourself

Search Engines

The Art of Searching

Suppose for a moment you search the recipe for Beef Stroganoff. (Just for an example. I know your mother's Beef Stroganoff is the best anyway.)
Were assuming you use the search engine Alta Vista in the example.

Probably your first thought would be to enter "Beef Stroganoff recipe" and hit "Search". You would find 578.377 matches, and learn all about the argentinian meat industry, get to know Igor Stroganoff, (janitor of the lavatories in the St. Petersburg Ermitáge), and find a fascinating recipe for steamed mealworms in white wine - until you finally disconnect and prepare the delicious tofu hamburger Stroganoff with catsup that you came across early in your five hour session.

First thing to try is enter

stroganoff

in the search engine. Why not "Beef Stroganoff" ? Well, some people call it Beef, but it's really called Boef and sometimes Filet, etc. etc ... So you see that Beef would unnecessarily narrow your options. Plus people who call it Beef are likely to use catsup also...

As a result you will probably get 3367 entries. Restaurant adds, the homepage of Igor Stroganoff (janitor of the lavatories in the St. Petersburg Ermitáge), and recipes for "Tofu Hamburger Stroganoff"

So the next step, instead of starting to read Igor's homepage, is to think of something that will always turn up in a recipe, but never in a homepage or in a restaurant add, and include this in your query.

A good choice would be salt. Now, if in Alta Vista you were looking for "stroganoff salt" you would get every page with stroganoff in it, plus every page with salt. To tell Alta Vista that both words must be on the page, search for

+stroganoff +salt

This will still include the tofu nerds. So to ensure you get only delicious Beef Stroganoff recipes, make this

+stroganoff +salt -healthy

;o)

the "-" dismisses every page that has the word "healthy" on it...
This is a very important option. It is very likely, for instance, that there exists a secret_plan.txt for a cloaking device for steam driven spaceships that several russian scientists currently work on. Among these are Dr. Igor Stroganoff (former janitor of the lavatories in the St. Petersburg Ermitáge), and his wife Danuta Salt-Stroganowa. Since they are confined  to a secret gulag, none of them is healthy. The interest in the scientific community has been immense and therefore the plans are mirrored on hundreds of internet servers. Unfortunately these are the first 400 matches you get for your recipe search. To get rid of them, you can add -secret_plan to your query.

Ok, this will not happen to you, but eg. there are some pea counting scientists who assemble dumb word lists that are actually mirrored everywhere.

Another important option is searching for an exact phrase. Figure you want to find out about the cloaking device and only know that a Dr. Stroganoff is involved. Now you will learn to hate Beef Stroganoff. To search only for the Dr. (former janitor of the lavatories in the St. Petersburg Ermitáge), enclose the name in quotation marks:

"Igor Stroganoff"

This tells Alta Vista to look for the exact phrase, including capitals.

These are the first things you should find out about any search engine:

• how to make a keyword mandatory ("+")
• how to exclude a keyword ("-")
• how to search an exact phrase.

 stroganoff AND salt AND NOT(Igor OR Danuta OR (cloaking AND device) OR healthy)
 

A final, very important tip: Don't try to find anything in your mother language unless it happens to be English...

The Search Engines

• HotBot

This is what I use regularly. Has the most up-to-date documents in its database. There is a feature to search for specific data (images, videos...) or limiting the search to specific domains (Asia, Europe...) just by point and click. To use a boolean query you have to explicitly choose that option in the drop down list.

• Alta Vista

The best all-purpose search engine. Very up-to-date and comprehensive. It uses the easy +/- syntax for inclusion / exclusion discussed above. Recently a very useful feature "Refine search" has been added. Enter a naive search (Beef Stroganoff recipe), click "Refine" and it will actualy make a list of all the subjects this leads to (like tofu, cloaking device, etc.) You can simply check phrases that you want to exclude from or include in your search.
 
• Inference Find

This is a meta search engine. It queries several search engines and combines the results with some reasoning behind it. This will find documents a single search engine may overlook, but it will also find unrelated (useless) documents.
 
• Lycos

A dinosaur. Has the most documents of any age in its database
 
• Yahoo!

Best if you have no keywords but like to browse categories. Note that this is not information research, but surfing :)
 
• DejaNews

This is a usenet archive. It archives the usenet discussion groups (newsgroups). You should search this if you don't seek general information, but answers. It is very likely that someone has asked your question in a newsgroup already. By default it searches only in new articles. Switch to "Past articles" if you are not researching something very recent, because these are many more articles.
Tip: to find the answers, and not only the desperate postings of guys who share your problem but never got an answer, include the "word" Re: in your query. This turns up in the header of any regular answer.

Archie "Gateways"

• ftpsearch.ntnu.no

This has by far the best "hit ratio"
• SWITCH ArchiePlexForm
• ArchiePlexForm

       fpArchie

It looks and feels just like the Win95 find files dialog, but it searches files on the internet instead of on your local disk. It also has built in FTP to download the files.

And what about you ?

Theres a number: 

in hoc signovincimus