Web searching

Introduction

• Quality of information
• Types of tools
• Search strategies
• Security

Quality control

• Nobody is in charge
• There are few reliable stamps of approval on the Web
• Information is not free if it could make money
• Things change very fast

New Yorker, 1993 Jul 5, p. 61

The authoritativeness of a page can be assessed in terms of

• qualifications of the author
• sponsoring organization
• potential conflicts of interest
• references to sources of information

Of course it's not just the Internet that can lie.
Some call-centre workers in India not only perfect American accents, but they claim American names and life stories if asked personal questions.
("Hi, I'm from Chicago", story by M. Landler, New York Times, seen in Montréal Gazette, 2001 Mar 22)

Subject directories

Subject directories provide human-generated hierarchical classifications of Web sites. These classifications are browsable (and maybe also searchable).

Subject directories do not include all (or even most) Web sites, just those which have somehow been identified as useful or relevant. There may be limited consistency in such evaluations. Explicit site ratings may be provided.

The selection of sites may be restricted to specific topics (e.g., health or medicine).

Some subject directories contain a lot of out-dated material and dead links (but then so do some search engines).

Examples of Subject directories

Search engines

Search engines provide the ability to search for text strings or words within large collections of text and addresses.
These collections are automatically generated by robots (or spiders) which mechanically traverse the Web and collect text.
The text is automatically converted by an indexing engine to a format which can be efficiently processed by a search engine.

Many search engines began as research projects and evolved into commercial operations including ‘portals’ and supported by advertising, paid services, etc.

There is strong competition and rapid change.

Examples of search engines

Effective searching

• Plan your search
• Know the tools
• Use the tools wisely
• Keep track of what you've done

Based in part on Effective Web Searching: an introductory slide show by Karen Campbell, 1997 March.

Planning a search

Traecy time rule: How much time is the answer worth?

• Decide before you start how much this is worth to you in terms of time.
• After spending that much time, revert to traditional sources or seek expert advice.

If searching via a slow Internet connection (e.g., modem and telephone line), consider turning off your browser's graphics. Or do the search later where you will have direct Internet access (e.g., at school).

Do you require a quick answer or a comprehensive answer?

• Decide whether you need something or everything
• If everything, try some meta-search engines
  • Duplicates may not be well organised
  • Different search engines used may interpret the search request in different ways

What other resources do you have?

Using the tools

Start with a subject directory unless you have very specific query.

Know the engines

• Find an engine (or engines) that works for you. Don't worry about which is “best”, find one which works well for you.
• Learn its address. Bookmark it in your browser, or add a link to your home page.
• Consider specialized search engines
• Don't just rely on your browser's built-in Search button.
• Read the help screens, and take advantage of your favourite engine's features.
• Look at the ‘advanced’ features, e.g. Boolean searches, proximity, etc.
• Keep up with their developments.
• Periodically try other engines – a new one may become your favourite.

Common methods

• Enclose phrases in quotes
• Think about synonyms

The following are less popular than they used to be:

• Put a + before all required words; a - before excluded words
• Capitalize proper nouns to avoid their occurrence as ordinary nouns
• Use truncation marks (usually * or ?)

Relevance ranking

Virtually all Web search engines attempt to rank hits by relevance. There are various methods for estimating relevance.

Look only at the first 20 or 30 hits to judge whether the search has succeeded.

If it hasn't, try different search words, or try a different search engine, or try a subject directory, or try Medline, or ...

A good search strategy for some search engines is to include several terms in your search, including synonyms.

Keep track of your results

• Keep a log of your search(es)
• Use bookmarks or a simple web page
• Find good resources and visit them often

Further information

• Pandia tutorial on Web searching, with comments and tips on various search engines
• Searching the Web in Yahoo!
• McGill Libraries' Internet Searching

Safe surfing

Remember the cautions about opening attachments, Word macros, etc.

Set the security level of your Web browser to refuse any ActiveX control or plug-in which is not signed by an entity you trust. ActiveX controls and plug-ins have complete access to your machine, unlike Java applets which are restricted to a sandbox (details).

Java (and Javascript) have also suffered from security problems in various versions of various browsers; use up-to-date software and the latest security patches, and/or disable Java and Javascript.