(This is originally a chapter from the book Efficient information searching on the web.)
Different types of information seeking
There are many types of information seeking. It could concern:
- Hunting for specific information, perhaps a single fact.
- Collecting of information to be forwarded, i.e., sending it on to somebody else.
- Tracing information for retrieval, i.e., a search for a known object, perhaps a Web page that you have visited earlier.
- Collecting of information for later reading and processing.
- Browsing information for learning, i.e., surfing around on subject Web sites or looking through these in a specific subject.
- Making an exhaustive search, i.e., getting all the information about a subject.
The mode of procedure when searching for factual information is not the same as when you collect information for later processing. Certain general modes of procedure exist. You always have two different ways to go about your search:
- Do a broad search and then narrow it down little by little.
- Do a narrow search and broaden it if required.
Broad searching may mean that you’re using few or general search words when you execute the search. Then you add more words or substitute general words for more specific words.
Narrow searching means to start out from several narrow and specific search words in the search. Then you remove words or you substitute specific words for general words.
Searching versus browsing
When searching on the Web there are two fundamentally different ways of searching. The first one is to follow links, i.e., to click yourself forward in your hunt for information. To follow links is often called surfing or browsing, depending upon how conscious the clicking is. Surfing is more random and unfocused than the browsing which is conscious and controlled.
The other way of searching is to search with words in a search service or within a Web site. Word searches are sometimes called analytical searching as it requires you to analyze your search subject and choose search words before you begin to search. When it comes to browsing you don’t have to have the search query formulated or chosen your search words, but can narrow down the sought information gradually.
Fig. Analytical strategies versus browsing strategies according to Gary Marchionini.
As the figure above illustrates, analytical strategies and browsing strategies differ in many different ways. The employment of the different search modes often depend upon the system in which you’re doing your search and what you’re searching for. We are no doubt also different as individuals and probably prefer different ways of searching.
The Booelan operators (AND and OR) occur in some part of the strategies below, they are further explained in Chapter 8.
The different concepts of the search query form search strings that are then combined. Traditionally, in databases one search has been executed for each block before combining them to a final search.
Search query: I want to find information about medical libraries or hospital libraries in the Swedish neighbouring cities Lund or Malmö.
The blocks might then look like this:
Block 1: lund OR malmö
Block 2: hospital library OR (hospital AND library)
Block 3: ”medical library”
In the final search the blocks are combined to a complete search string and a parenthesis is used for a correct combination of the blocks:
Final search: (lund OR malmö) AND (hospital library OR [hospital AND library] OR ”medical library”)
If a search consists of two or various separate terms the different parts are called facets. Within each facet the search words may be supplemented with synonyms and other closely related terms. Everything to get as good coverage as possible for the facet. The different facets are then combined in a joint search.
If you’re looking for information about transport costs the facets will be transport and costs. The transport facet can be increased by freight and ”shipment of goods” (one phrase, must be marked with quotation marks). The terms should be combined with OR so that only one of the search words is required to get a hit. Costs can in the other facet be supplemented with price. The two facets are combined with AND. The search string will be:
(transport OR freight OR ”shipment of goods”) AND (cost OR costs OR price)
A variant of the facet search is ”successive fractions”. This means that the original, big search block is narrowed down little by little with increasingly finer limitations by adding more narrow concepts.
You can also start with the most specific concept. In that way you can work more efficiently by not having to do part of the work with fractions as you start out with a smaller amount.
Pearl growing means that you pull out search words from a relevant document, e.g., through citing a name or subject word. Through the search concepts that have been fished out (the pearls) new documents are found that have a connection with the first one (more pearls).
If you have a good reference to start out from, e.g., a good Web page, you can grow this information. In a search engine you can often do a search with the operator link: to get the Web pages that have linked to the page of interest. The Web pages that have linked to your resource page may contain relevant information or link to other similar pages.
Internetbrus.com is a Swedish Web blog about searching and search services on the Internet. In Google you can search for the pages that link to this very page by writing
In Google you shouldn’t enter http://, but www should be entered if the address includes this.
You can do the same with books in a Net bookstore. If you find your book you can often see what other books customers of your chosen book have bought; you might get lucky. In Amazon (www.amazon.com), for example, you can also search on the categories that the book belongs to.
Interactive scanning means that you search through a great number of relevant documents for key concepts to formulate new problems, which may then result in new search queries.
Simple searching (quick-and-dirty)
A simple search is a search on one or a couple of search words combined with AND or OR. No subtleties are used and the result, in any case in databases, will be rough but it will provide a first image. Sometimes the term quick-and-dirty is used by “pro searchers” for simple searches as the very art of searching just because it’s fast, but the result is a bit messy. Almost all searches on the Web are simple searches.
The browsing strategies are important ways to search the Web as they can make use of the Web’s nature of using links. Traditionally, browsing hasn’t been considered as professional as analytical searching, in part owing to the connections that the companies in the search business have with the database industry. Perhaps it’s not until now, with the Web and the new generation of services sometimes called Web 2.0, that browsing may be seen as equal in merit to analytical searching.
Specific (controlled) browsing is systematic, focused on a specific object or goal, e.g., the scanning of a list of known objects to verify facts.
Predictive (semi-controlled) browsing is generally purposeful, the goal is less delimited or clear and the browsing is less systematic.
General (uncontrolled) browsing doesn’t have a real goal and very little focus, e.g., flipping through a magazine and zapping the TV.
Berrypicking is, broadly speaking, a browsing strategy, but with elements of analytical searching. It’s really a searching behaviour which most of us have and that has been formalized and named. Berrypicking means to collect, or pick, information little by little as you find it, like other browsing, but you actually do it irrespective of search system or search strategy. The search is done in the same way as berry picking in the woods, unplanned but still structured and adapted to the surroundings.
Other search strategies
Besides searching with words or browsing there are a few other ways to search:
Enter the URL
To enter a Web address requires previous knowledge of the address or an idea of where the sought information may be found. If you know that all Swedish municipalities have URLs set up according to the same pattern (www.nameofmuniciaplity.se) it’s not so hard to figure out that Ronneby municipality is found at www.ronneby.se and Stockholm municipality at www.stockholm.se. Likewise you may conclude that there will be quite a bit of information about space at www.nasa.gov.
Monitoring a subject
If you have a constant interest in a subject there are several ways to monitor what’s going on. Different variants are:
Standing search in a search engine (alert) – you receive a mail when there are new matches for your search terms.
E-mail lists – through membership in an e-mail list information about the list theme is sent to your inbox. Companies or interest groups may have newsletters and in this way you’ll be getting new information. Or the list may be intended for like-minded people who can send letters (and reply) to the list.
Two-step searching is an indirect search, e.g., to rapidly find sought information or to find information which can’t be found through the search engines. The search is done in various steps (at least two). You have to do the search in two steps, first find the place which holds the information and then look in that place. As far as databases are concerned, the search page is often indexed in the search engines and can thereby be found.
Three different two-step searches as examples:
Middle Ages links: site:www.museumof london.org.uk middle ages links (in Google or other search engine depending on how site site: is handled).
Databases about airplane crashes : database “plane crash” (in a search engine).
Music by Apoptygma Berzerk for listening: listen “apoptygma berzerk” (in search engine).
Two-step searching is the most important search strategy on the Invisible Web, see Chapter 11.
Subject-specific words – lingo
For each search there is a vocabulary, a slang word or other unique words that make it easier to get relevant hits.
Other interested parties
For most subjects there are at least a few interested people, and some of them have already put up a Web site, a forum or other resources on the Net in the subject where they collect and comment on relevant links. You can facilitate your search a lot by using these special resources instead of trying to build your own.
From the inside and out
When you search it’s better to start with a very specific search query and then get increasingly more general. If you start with more general queries you will be flooded with hits. If the search doesn’t work, make it slowly more general until you find a good balance between the number of hits and the contents of the hits. – More is not better!
A loved child has many names
Most things, unfortunatly, have nicknames. New York can be NY, Carl-Gustav can be CG or Calle. Nicknames have to be included in an exhaustive search.
Just the right puzzle piece
You never know which piece of information will be the one which brings your entire search together. So keep your eyes open, look for each fragment of information that has the potential to make your search complete.
Don’t forget your fellow human being
When you’re searching on the Net it’s important to remember that there is a world beyond the Internet. Web pages are created by human beings and sometimes human beings are the most important resource you can find.
The Web is constantly expanding, constantly changing, constantly breaking down (unfortunately), and constantly rebuilt. If you’re planning to do continuing investigations it’s important to have a strategy to keep informed about all the new Web sites and resources in your subject. (See Chapters 12 and 15)
Increase your search power
If you use the Web site’s special syntax you’ll increase the power in your searches considerably. If you mix the special syntaxes you’ll increase the power in your searches still more. But be careful. (See Chapter 8)
You should take each page on the Internet with a grain of salt. (See Chapter 9 )
Many roads lead to the same place
On the Web there are often several ways to reach the intended destination. If it’s impossible to get ahead on the route you planned, you have to find another way to reach your goal.
The speed and range of the Internet is awfully seductive, but remember that most of humanity’s saved knowledge is stored in non-electronic, durable media. Written or printed text is the biggest source, but information may also be extracted from sculptures, monuments and other cultural objects.
Go directly to the source
If possible, you should go directly to the source. If there are authorities, institutions or organizations connected to the subject area you’re interested in, you’ll probably find links on their Web sites. Many of these have librarians or Web editors who work professionally with information in the area.
Marchionini, Gary, Information Seeking in Electronic Environments, 1995.
 Bates, Marcia J. “The Design of Browsing and Berrypicking Techniques for the Online Search Interface.” Online Review 13 (October 1989): 407-424. [http://www.gseis.ucla.edu/faculty/bates/berrypicking.html]