THE POSSIBILITIES PROLIFERATE, CONSIDER ADOPTING A QUERY-BASED APPROACH
Teaching how to search the Web is hot--it's become standard practice in public-service librarianship. We teach classes, hold workshops, distribute handouts, mount tutorials on the Web. We know our sources and don't hesitate to help. Do you want a list of the top search engines? The most substantive directories? Good places to find deep Web content? No problem!
Still, basic questions remain. How are we going about teaching search tools? Are users taking away information they can really use? Does our instruction hold up over time?
From what I've seen, I have concluded that many of us take a descriptive approach to teaching the Web. We are masters at identifying search tools and describing their salient characteristics. We routinely detail their query options, search syntax, and results-ranking schemes. Our instruction is loaded with facts.
The answer lies in the question
As admirable as this approach is, it is ultimately unproductive. Description isn't much help in an environment in which search tools and their features are madly proliferating on an ever-burgeoning Web. There is simply too much to remember, and too much change. How many of us have expended good time and effort giving an in-depth review of a search tool's features, only to see these features mutate or disappear? This is frustrating to us, and it leaves our users with little that they can use.
I propose that we take an approach that is more apt to provide knowledge that endures. I call this a query-based approach because it is based on the individual query. If we teach search tools for their usefulness with specific types of queries, we are giving users a reason to return to these tools. We are identifying needs and finding solutions. Each tool becomes not just a bundle of characteristics but something that fits into an information-finding context. In short, we are providing users with a strategy.
Here are 10 tips for shaping a query-based approach.
1. Apply what we know about library resources to search tools on the Web. Librarians understand the characteristics of library-based research tools and when it is appropriate to use them. With library resources, it is a matter of routine to match the query to the tool. We need to approach Web-based tools in the same spirit. When teaching how to do research on the Web, we should evaluate its information-finding tools based on the queries they support, and recommend them accordingly.
2. Help users find what they need. Users tend to view Web search tools as an amorphous, undifferentiated whole. By the same token, many users believe that there are tools that can answer all queries. If they try a certain tool and are disappointed, they tend to fault the tool rather than their decision to use that tool. It's our job to explain that different search tools serve different purposes and to help users identify which tools match their needs.
3. Define searching broadly. To teach query-based searching, we need to view searching as a process that begins with the quest for information. It is far more than the act of constructing a search statement--in fact, that step comes last. Users should begin the search process by analyzing their query. Do they want to begin with a broad topic and become familiar with its subtopics? Is their topic targeted to a narrow concept or made up of multiple concepts? Are they looking for a specific Web site? Do they want a targeted set of data? Dynamically changing information? Multimedia?
Based on the answers to these kinds of questions, we can help users explore the tools that might bring them results. Once they have chosen the right tool, we can address the matter of constructing search statements.
4. Teach search tools, not just search engines. The Web offers three major types of information-finding tools: directories, search engines, and the deep Web. Teaching search engines alone is not enough. A query-based approach to Web searching encompasses all types of Web-based tools. We need to familiarize users with the full range of tools and the kinds of queries they can address. Search-engine training is limiting, while search-tool training opens up a world of possibility.
5. Teach users to analyze their queries and identify the tools that support them. The following illustrates how useful a query-based approach can be. Here are three queries, all on the topic of American architecture, each of which requires a different type of Web-based tool.
I'm looking for: sites on American architecture.
Use: A professional directory created and annotated by experts.
I'm looking for: the site of the Society of American Registered Architects.
Use: A peer-ranking, human-mediated engine such as Google or Direct Hit.
I'm looking for: a list of architects in Baltimore.
Use: A database on the deep Web such as a phone book.
These examples demonstrate the advantage of analyzing the query first, then choosing the search tool as a second step. Tips 6, 7, and 8 cover each major type of tool in greater detail.
6. For general queries or for topics that need exploring, recommend directories. Directories, especially those compiled and annotated by experts, are appropriate starting points for broad topics. A few examples are the Argus Clearinghouse, InfoMine, and the Librarians' Index to the Internet. These tools give users an opportunity to see what the experts have to say about the best resources available on their topic. With their hierarchical subject listings, directories are also good for browsing. Listed subtopics can help users become familiar with the scope of their topic for further refinement. In addition, directories often include meta pages that are jumping-off points for topical research. It's important to teach directories as human-mediated tools that tend to offer substantive content.
7. For targeted, ambiguous, and sometimes broad queries, recommend search engines. Traditionally, search engines have worked best for targeted or multi-concept queries. Because we are searching the full text of millions of files, we are able to pick up specific and often obscure information. With the current crop of engines, an even wider range of queries is supported. The following examples illustrate this point. These queries range from the specific to the very broad.
Query type: targeted to a narrow topic.
Query: I'd like to view sites about the Hubble telescope.
Use: Peer-ranking, human-mediated engines.
Examples: Google, Direct Hit.
Why? The Web is a community of content creators and users of this content. People who link to external sites from their Web pages exercise judgment about the relevance and value of these sites. Google's relevancy ranking measures this activity. Direct Hit tracks the sites that users select from their search engine results. The collective judgment of millions of these searchers adds up to a continual and dynamic peer ranking. Both types of rankings work quite well when we are searching a narrowly-defined topic.
Query type: targeted to a specific site or other restriction.
Query: I'd like to view NASA documents about the Hubble telescope.
Use: Engines with a searchable site field.
Examples: AllTheWeb, AltaVista, HotBot, IxQuick Metasearch, Northern Light.
Why? Engines that offer "site" or "URL" as a field restriction allow us to retrieve documents from a specific site. These limits may be put into effect through search syntax or menu choices in a search template. This idea can be extended to other types of field delimiters such as geographic location ("I want to see documents from South Africa about Nelson Mandela"), date last modified, language, file type, etc. A number of search engines work well for these types of targeted queries.
Query type: ambiguous or terminology-seeking.
Query: I'm interested in learning about stocks.
Use: Concept-processing, thesaurus-creating engines.
Examples: Excite, SurfWax.
Why? Ambiguous words are always a challenge in a database search. Thesaurus-creating engines can help us narrow our concept to our intended meaning. These engines offer a choice of meanings based on the initial search, from which users can select for a subsequent search. Thesaurus-creating engines, like their library-based counterparts, can also help users choose appropriate terminology for a search.
Query type: general, in-depth.
Query: I'm doing research on renewable energy.
Use: Concept-clustering tools that parse topics into component subtopics.
Examples: Northern Light, Guidebeam, Query Server, Vivisimo.
Why? Concept-clustering tools process a search and return results that are organized into subtopics and relevant sites. This can be very useful when you want to become familiar with different aspects of a topic, are unfamiliar with a topic, or want to be sure you are examining it in depth. In this respect, these tools and directories serve a similar purpose.
8. For information stored in databases or non-textual files, recommend the deep Web. Fixed Web pages are only one part of the content available on the Web. The much larger part is held in databases or nontextual files. Data, graphics, software, dynamically changing information, and multimedia are examples of deep Web content.
This content may be retrieved in a variety of ways. Many databases on the Web are searchable from their own sites, and these sites can be retrieved from directories and search engines. Also, many search engines offer deep Web searches as featured options. For example, it is not unusual to be offered searches for news, multimedia, stock prices, airline tickets, items in Web stores, and much more. A few sites specialize in gathering a collection of links to searchable databases on the Web, for example the Invisible Web. Others, such as ProFusion, search the content of selected databases from a single interface.
9. Avoid getting bogged down in teaching search-tool features. Features come and features go. Trying to keep track of which ones belong with which tool is very difficult. If this is a challenge for us, what about our users? Even if we could keep track of everything, teaching features in and of themselves has little value. We should avoid an approach that says, "This tool does this, that tool does that." This leaves users with numerous details but no grounds for using the tool once they're on their own. It's much better to say, "Search engines have features and they change." Then, give advice about the features to look for based on the nature of the query. Remember: context is everything.
10. Be realistic--and relax! It's amazing to think that we are still in the early years of information-finding tools on the Web. The volatility of this world is sure to continue. Absolve yourself and your users of the burden of tracking a multiplicity of details. Instead, teach what is useful in the actual process of finding information. Rather than elaborating on features, put your attention on the query. This is a lesson that will stand the test of time. degrees7degrees
Steps for this search and those to come:
• Define the nature of the quest.
• Choose the most useful tool.
• Construct the proper search statement.
SEARCH TOOLS AND THEIR URLS
Google: http://www.google.com/
AllTheWeb: alltheweb.com
AltaVista: http://www.altavista.com/
Argus Clearinghouse: http://www.clearinghouse.net/
Direct Hit: http://www.directhit.com/
Excite: http://www.excite.com/
Guidebeam: guidebeam.com
HotBot: hotbot.lycos.com
InfoMine: http://www.infomine.com/
Invisible Web: http://www.invisibleweb.com/
IxQuick Metasearch: http://www.ixquick.com/
Librarians' Index to the Internet: lii.org
Northern Light: http://www.northernlight.com/
ProFusion: http://www.profusion.com/
Query Server: http://www.queryserver.com/
SurfWax: http://www.surfwax.com/
Vivisimo: vivisimo.com
By: Cohen, Laura B., American Libraries
Teaching how to search the Web is hot--it's become standard practice in public-service librarianship. We teach classes, hold workshops, distribute handouts, mount tutorials on the Web. We know our sources and don't hesitate to help. Do you want a list of the top search engines? The most substantive directories? Good places to find deep Web content? No problem!
Still, basic questions remain. How are we going about teaching search tools? Are users taking away information they can really use? Does our instruction hold up over time?
From what I've seen, I have concluded that many of us take a descriptive approach to teaching the Web. We are masters at identifying search tools and describing their salient characteristics. We routinely detail their query options, search syntax, and results-ranking schemes. Our instruction is loaded with facts.
The answer lies in the question
As admirable as this approach is, it is ultimately unproductive. Description isn't much help in an environment in which search tools and their features are madly proliferating on an ever-burgeoning Web. There is simply too much to remember, and too much change. How many of us have expended good time and effort giving an in-depth review of a search tool's features, only to see these features mutate or disappear? This is frustrating to us, and it leaves our users with little that they can use.
I propose that we take an approach that is more apt to provide knowledge that endures. I call this a query-based approach because it is based on the individual query. If we teach search tools for their usefulness with specific types of queries, we are giving users a reason to return to these tools. We are identifying needs and finding solutions. Each tool becomes not just a bundle of characteristics but something that fits into an information-finding context. In short, we are providing users with a strategy.
Here are 10 tips for shaping a query-based approach.
1. Apply what we know about library resources to search tools on the Web. Librarians understand the characteristics of library-based research tools and when it is appropriate to use them. With library resources, it is a matter of routine to match the query to the tool. We need to approach Web-based tools in the same spirit. When teaching how to do research on the Web, we should evaluate its information-finding tools based on the queries they support, and recommend them accordingly.
2. Help users find what they need. Users tend to view Web search tools as an amorphous, undifferentiated whole. By the same token, many users believe that there are tools that can answer all queries. If they try a certain tool and are disappointed, they tend to fault the tool rather than their decision to use that tool. It's our job to explain that different search tools serve different purposes and to help users identify which tools match their needs.
3. Define searching broadly. To teach query-based searching, we need to view searching as a process that begins with the quest for information. It is far more than the act of constructing a search statement--in fact, that step comes last. Users should begin the search process by analyzing their query. Do they want to begin with a broad topic and become familiar with its subtopics? Is their topic targeted to a narrow concept or made up of multiple concepts? Are they looking for a specific Web site? Do they want a targeted set of data? Dynamically changing information? Multimedia?
Based on the answers to these kinds of questions, we can help users explore the tools that might bring them results. Once they have chosen the right tool, we can address the matter of constructing search statements.
4. Teach search tools, not just search engines. The Web offers three major types of information-finding tools: directories, search engines, and the deep Web. Teaching search engines alone is not enough. A query-based approach to Web searching encompasses all types of Web-based tools. We need to familiarize users with the full range of tools and the kinds of queries they can address. Search-engine training is limiting, while search-tool training opens up a world of possibility.
5. Teach users to analyze their queries and identify the tools that support them. The following illustrates how useful a query-based approach can be. Here are three queries, all on the topic of American architecture, each of which requires a different type of Web-based tool.
I'm looking for: sites on American architecture.
Use: A professional directory created and annotated by experts.
I'm looking for: the site of the Society of American Registered Architects.
Use: A peer-ranking, human-mediated engine such as Google or Direct Hit.
I'm looking for: a list of architects in Baltimore.
Use: A database on the deep Web such as a phone book.
These examples demonstrate the advantage of analyzing the query first, then choosing the search tool as a second step. Tips 6, 7, and 8 cover each major type of tool in greater detail.
6. For general queries or for topics that need exploring, recommend directories. Directories, especially those compiled and annotated by experts, are appropriate starting points for broad topics. A few examples are the Argus Clearinghouse, InfoMine, and the Librarians' Index to the Internet. These tools give users an opportunity to see what the experts have to say about the best resources available on their topic. With their hierarchical subject listings, directories are also good for browsing. Listed subtopics can help users become familiar with the scope of their topic for further refinement. In addition, directories often include meta pages that are jumping-off points for topical research. It's important to teach directories as human-mediated tools that tend to offer substantive content.
7. For targeted, ambiguous, and sometimes broad queries, recommend search engines. Traditionally, search engines have worked best for targeted or multi-concept queries. Because we are searching the full text of millions of files, we are able to pick up specific and often obscure information. With the current crop of engines, an even wider range of queries is supported. The following examples illustrate this point. These queries range from the specific to the very broad.
Query type: targeted to a narrow topic.
Query: I'd like to view sites about the Hubble telescope.
Use: Peer-ranking, human-mediated engines.
Examples: Google, Direct Hit.
Why? The Web is a community of content creators and users of this content. People who link to external sites from their Web pages exercise judgment about the relevance and value of these sites. Google's relevancy ranking measures this activity. Direct Hit tracks the sites that users select from their search engine results. The collective judgment of millions of these searchers adds up to a continual and dynamic peer ranking. Both types of rankings work quite well when we are searching a narrowly-defined topic.
Query type: targeted to a specific site or other restriction.
Query: I'd like to view NASA documents about the Hubble telescope.
Use: Engines with a searchable site field.
Examples: AllTheWeb, AltaVista, HotBot, IxQuick Metasearch, Northern Light.
Why? Engines that offer "site" or "URL" as a field restriction allow us to retrieve documents from a specific site. These limits may be put into effect through search syntax or menu choices in a search template. This idea can be extended to other types of field delimiters such as geographic location ("I want to see documents from South Africa about Nelson Mandela"), date last modified, language, file type, etc. A number of search engines work well for these types of targeted queries.
Query type: ambiguous or terminology-seeking.
Query: I'm interested in learning about stocks.
Use: Concept-processing, thesaurus-creating engines.
Examples: Excite, SurfWax.
Why? Ambiguous words are always a challenge in a database search. Thesaurus-creating engines can help us narrow our concept to our intended meaning. These engines offer a choice of meanings based on the initial search, from which users can select for a subsequent search. Thesaurus-creating engines, like their library-based counterparts, can also help users choose appropriate terminology for a search.
Query type: general, in-depth.
Query: I'm doing research on renewable energy.
Use: Concept-clustering tools that parse topics into component subtopics.
Examples: Northern Light, Guidebeam, Query Server, Vivisimo.
Why? Concept-clustering tools process a search and return results that are organized into subtopics and relevant sites. This can be very useful when you want to become familiar with different aspects of a topic, are unfamiliar with a topic, or want to be sure you are examining it in depth. In this respect, these tools and directories serve a similar purpose.
8. For information stored in databases or non-textual files, recommend the deep Web. Fixed Web pages are only one part of the content available on the Web. The much larger part is held in databases or nontextual files. Data, graphics, software, dynamically changing information, and multimedia are examples of deep Web content.
This content may be retrieved in a variety of ways. Many databases on the Web are searchable from their own sites, and these sites can be retrieved from directories and search engines. Also, many search engines offer deep Web searches as featured options. For example, it is not unusual to be offered searches for news, multimedia, stock prices, airline tickets, items in Web stores, and much more. A few sites specialize in gathering a collection of links to searchable databases on the Web, for example the Invisible Web. Others, such as ProFusion, search the content of selected databases from a single interface.
9. Avoid getting bogged down in teaching search-tool features. Features come and features go. Trying to keep track of which ones belong with which tool is very difficult. If this is a challenge for us, what about our users? Even if we could keep track of everything, teaching features in and of themselves has little value. We should avoid an approach that says, "This tool does this, that tool does that." This leaves users with numerous details but no grounds for using the tool once they're on their own. It's much better to say, "Search engines have features and they change." Then, give advice about the features to look for based on the nature of the query. Remember: context is everything.
10. Be realistic--and relax! It's amazing to think that we are still in the early years of information-finding tools on the Web. The volatility of this world is sure to continue. Absolve yourself and your users of the burden of tracking a multiplicity of details. Instead, teach what is useful in the actual process of finding information. Rather than elaborating on features, put your attention on the query. This is a lesson that will stand the test of time. degrees7degrees
Steps for this search and those to come:
• Define the nature of the quest.
• Choose the most useful tool.
• Construct the proper search statement.
SEARCH TOOLS AND THEIR URLS
Google: http://www.google.com/
AllTheWeb: alltheweb.com
AltaVista: http://www.altavista.com/
Argus Clearinghouse: http://www.clearinghouse.net/
Direct Hit: http://www.directhit.com/
Excite: http://www.excite.com/
Guidebeam: guidebeam.com
HotBot: hotbot.lycos.com
InfoMine: http://www.infomine.com/
Invisible Web: http://www.invisibleweb.com/
IxQuick Metasearch: http://www.ixquick.com/
Librarians' Index to the Internet: lii.org
Northern Light: http://www.northernlight.com/
ProFusion: http://www.profusion.com/
Query Server: http://www.queryserver.com/
SurfWax: http://www.surfwax.com/
Vivisimo: vivisimo.com
By: Cohen, Laura B., American Libraries
