Search Strategy, Search Engines

Search Strategy, Search Engines



Finding Information on the Internet

Options:


Searching the WWW for Information


Strategizing Process

"How do I find . . ." is the area where the WWW is weakest. A compounding problem is that plenty of information is not yet available on the Internet. Therefore, in order to find information in the most quick and efficient way possible, you need to develop strategies that will increase your chances of success.

Tip: Searching for a known item (URL,company/individual/organization name) offers the most chance for success.

Step #1: Think of likely sources that would offer the information you are seeking, ie. extrapolate what is probable from what is known.

Step #2: Which search method will work best?


Address Guessing

Now that you understand the anatomy of an address, can you use that knowledge to construct a URL for the source that you think might have the information you seek.

In this case, http://www.microsoft.com

Note: Often acronyms are used rather than full names. For instance, http://www.ibm.com

Important: Address guessing not 100% certain, so what other type of known-item search can you do. In other words, where do you go from here?

Answer: Net Search


Mailing lists

The Liszt Directory of Email Discussion Groups is at http://www.liszt.com/
And, you can do a keyword search for lists at http://catalog.com/vivian/interest-group-search.html, Search the List of Lists.

You are encouraged to join a mailing list on a topic of interest to you. By being in touch with people who have a similar interest as yours, you have access to a vast amount of experience and advice. If you join a mailing list, you are not expected to jump right into the conversation. "Lurking" is a time-honored tradition in mailing lists and newsgroups. Lurking means reading the messages that are generated from the list, but not contributing any of your own. In fact, it's strongly suggested that you lurk for at least a short while to get the flavor of the group, and not embarass yourself by saying something off-topic.


Search Engines (Keyword Searching)

Most effective for known-item searches. Not appropriate for subject searching - you'll get back an overwhelming number of returns.


How Keyword Searching Works

Robots (artificially intelligent agents) search the Internet, noting the text of Web documents, and building databases of words found in those documents. When you make a request, the search software builds its results from that database based on the words you input. Results can include people’s personal pages, as well as substantive information.


What Makes One Search Engine Different From Another?

All have different programs that use different criteria to build their databases with the intention to balance number of returns against precision. Consider the following list of selected Search Engines:

AltaVistaSearch tips:
  • Offers Simple and Advanced query modes.
  • In Simple query mode, use + in front of a word (no space) to require it, use - in front of a word (no space) to exclude it. Truncation symbol is * (e.g., child* searches for childn, childish, children). Use quotation marks for phrase searching (e.g. "cost of living"). A capital letter forces an exact match for a word beginning with that letter.
  • In Advanced query mode, Boolean operators AND, OR, AND NOT, and NEAR are supported.
  • See AltaVista Help files for additional ways to refine searches.
Results are ranked by how near search words are to the top of the document, by how close the search words are to one another, and by their number in the retrieved document.
ExciteSearch tips:
  • Uses "concept seraching"
  • Use quotation marks to search for phrases
  • Use + immediately in front of a word to require its use; use - immediately in front of a word to exclude it.
  • Supports Boolean context (Boolean operators MUST be capitalized
  • Go to Power Search for a forms-based aid to creating Boolean searches
Results are ranked by "confidence level" which calculates how closely the site matches the search request.
HotBot Search tips:
  • Popup menu allows searching for all the words, any of the words, the exact phrase, words in the title, the person, links to this URL, the Boolean expression.
  • HotBot is case insensitive, except for mixed upper and lower case.
  • Searches can be limited in a variety of ways (e.g., for images, by date).
  • To explore the options, click on OPEN ALL.
Results are listed by confidence ranking based on word frequency in document, search words in title, search words in keywords lists, and document length.
Infoseek Search tips:
  • Capitalize proper names
  • Use quotation marks around phrases or use hyphens between words to locate those words next to each other.
  • Use + immediately in front of a word to require its use; use - immediately in front of a word to exlude it.
  • To constrain searches to specific Web document fields, type the field name (either link, site, url or title, in lower case) followed by a colon and the word to be searched.
Results list provides title of Website, URL, score, size of file, and brief summary.
Lycos Search tips:
  • Use quotation marks around phrases.
  • Use + before a word to require it; use - before a word to exlude it.
  • Click on Lycos Pro for Boolean and natural language queries.
Results are ranked by decreasing relevance based on word frequency counts. Three display options: summary, standard, and detailed. User can also select the number of results to be presented.


Meta-search Tools

Use these tools to simultaneously search multiple search engines from the same screen.
Dogpile Dogpile is a multi-engine search tool. It can limit the search to the Web, newswires, newsgroups, or FTP sites. About 15 standard search engines are searched simultaneously.
MetaCrawler MetaCrawler provides a single interface for parallel searches of Lycos, InfoSeek, and four other search engines. It can sort results in a variety of ways.
Savvy Search Savy Search provides one-screen access to about fifteen search engines. Results can be displayed in order by search engine, or can be integrated into one list.


Subject Lists and Directories

Web directories make excellent resources and, if you know how to use them, will help you find almost anything you want. The people who maintain these directories use special programs to do most of the work. These programs (called spiders, worms, crawlers or robots) search the Web looking for new and updated items.

There are a number of free, well-maintained directories available on the Web and, in general, there are two ways to use them. First, some directories have home pages organized into categories. To use such a directory, you access the home page and choose the category you want. Within that category, there may be subcategories. Thus you can select categories and subcategories, until you narrow down your search to find exactly what you want.

Other directories allow you to look for something by using a search engine. To use a search engine, you use your browser to fill in a form in which you describe what you are looking for. Once you submit the form, it is sent to the search engine which looks through the directory, finds what you want (if it exists) and sends back the results.

The following tools provide subject lists and directory lists to simplify searches.

Yahoo A directory of Internet links grouped by subject categories and subcategories. Click on Options for enhanced search capabilities.
Argus Clearinghouse Topical guides compiled by librarians that identify, describe, and evaluate Internet-based information sources on a wide range of topics.
Britannica Internet Guide Extensive guide to selected Internet resources organized by topical area, from Encyclopedia Britannica.
BUBL Link The Bulletin Board for Libraries (BUBL) organizes access to new Internet resources by subject categories.
Librarians' Index to the Internet Extensive guide to Internet resources by category; can also search by keywords.
WWW Virtual Library Subject experts from around the world contribute to this collection of subject oriented guides to the Internet.


Situation:

You want to buy a car. You don't mind where you buy it, if a reduced sales tax in Santa Cruz or another nearby county will save you money.

  1. Step #1: Make your situation into a question: What are the county sales tax rates in California?

  2. Step #2:
    • Think of likely sources that would offer the information you are seeking.
      Answer: California State Board of Equalization

    • If you can't think of a likely source, take keywords from your question.
      Answer: county sales tax rate california

  3. Step #3: Choose a search engine
    • HotBot - searches every page in its entirety once a week. Allows you to refine your search criteria considerably.
    • Excite - uses "concept searching" of returned documents to refine keywords and criteria (click on "more of these")
    • SavvySearch
      Special features:
      • Service tries to decide which 2 or 3 engines will work best for your search.
      • List of engines includes those that search gopher, such as WebCrawler and Open Text
      • Can search in different languages.

  4. Step #4: Type in your keywords. Choose your criteria carefully, if you are given that choice.

  5. Step #5: When results appear on the screen, look at the title, description and address of each hit to help you decide which is a possible match.

  6. Step #6: When you link into a page, if the information you are seeking is not immediately apparent, use "Find" button.

  7. Step #7: Check the timeliness of the data and who is providing it.


Suggestions:

  1. Enter words or phrases that are likely to appear in the document you want to find. Tip: Technical terms will yield more specific results than frequently used words (Example: columbine vs flower).

  2. Be flexible with your terms. Think of synonyms and variations on words. Use general library and/or online reference resources, such as encyclopedias, dictionaries, and thesauri to better understand your topic.

  3. Avoid overloading search engine - 3 to 5 words is best.

  4. Type the most important word/phrase first.

  5. Check spelling.

  6. If the search engine offers options, you may want to change some of the default settings to better control the results of a search. For instance, HotBot gives you the option of matching just some or all the terms that you type in. If you want to learn more about how to effectively search with any of the engines, they all have a "Hints" link.

  7. If you aren’t satisifed with the results returned by the search engine, try a different one. Each search engine uses different logical operators, therefore each will "think" differently.

  8. If too many results are returned, add another keyword.


Back to Top

Internet Search Exercises


[Send Email to Instructor] [Internet Resources: CIS/LBT 160] [ ] [Introduction, Web Browsers, Navigation, Email] [Browser and Email Exercises] [Search Strategy, Search Engines] [Internet Search Exercises] [Finding Files & Finding People] [Finding Places and People Exercises] [Mailing Lists and Newsgroups] [Mailing List Exercises] [Gopher, Archie, Veronica, Jughead] [Gopher Exercises] [FTP: File Transfer Protocol] [FTP Exercises] [Telnet: Logging in to a remote computer] [Telnet Exercises] [Privacy, Security, Legal & Ethical Issues] [Ethical Issues: Exercises] [Evaluating Information Found on the WWW] [Evaluation Exercises]