2-3b-Holans

							INTERNET SEARCHING

  VLIR Workshop on Library
    Development Problems
   CBA Heverlee 13.2.2003
        Ludo Holans
                Lay-out
• Introduction
• Meta search engines
• Pitfalls
• Classic searching versus free internet
  searching
• New developments : hyperbolic browsers
  and the semantic web
                Introduction

• The majority of information that can be searched
  on the web is text-based.
• Searching techniques and operators
• a. Boolean operators
• b. Proximity search (near near/n with adj
•    before after)
• c. Wildcards (truncation)
• d. Nesting (parentheses) « »
• e. Field searching
  (domain,image,link,text,title,url)
   META SEARCH ENGINES
• A. Searching instruments
• 1. Directories (databases compiled and
  indexed by people). Hierarchical system of
  categories. Tree structure.
  http://www.britannica.com
• http://dmoz.org
• http://www.yahoo.com/
      A. Searching instruments

• 2. Search engines
•    Use of a spider or crawler in order to find new
  or renewed information. This information is than
  put in the database. The use of a stoplist can
  eliminate a great percentage of noise. Stemming
  (reduction of the keyword by its stem) enables the
  future use of wildcards.
• Examples : http://www.altavista.com/
•             http://www.excite.com/
•             http://www.google.com/
      A. Searching instruments
• 3. Metasearch instruments
•    a. Metasites bring searching instruments via
  hyperlinks of searching interfaces together on 1
  website or webpage.
• Examples :
• http://www.gogettem.com/
•
  http://stommel.tamu.edu/~baum/linuxlist/linuxlist/
  node4.html
     A. Searching instruments
• b. « Human resource » websites
• Via een form you send your query to a team
  of specialists. The results of your question
  are sent afterwards.
• Example : http://www.about.com/
       A. Searching instruments
•   c. Metasearch engines
•   Examples :
•   http://www.ixquick.com/
•   http://www.mamma.com/
•   http://www.MetaCrawler.com/
•   http://www.Profusion.com/
•   http://www.Queryserver.com/
•   http://www.search.com/
     B. Working of search engine
•   Matching of document features to a query
•   1. Term frequency
•   2. Location of terms
•   3. Link analysis
•   4. Popularity
•   5. Date of publication
•   6. Length
•   7. Proximity of query terms
•   8. Proper nouns
               PITFALLS
• Common errors :
• 1. Mispellings (searching, serching,
  seerching, sherching etc)
• 2. Redundant terms (6 to 8 terms)
• 3. Stoplist terms
• 4. Too many terms, synonyms
• 5. Alternate spellings
               PITFALLS
• 6. Alternate spellings
• 7. Construction too complicated or not the
  correct Boolean construction

• USING FILTERS for slicing and dicing
  your search results
                PITFALLS
• Filters
• a. Site filters
•    com, edu, gov, mil, net, org, arts, firm,
  info, nom, rec, store, web
•    country domains (geo or ISO3166
  domains)
• b. Size and data filters
•
                    Filters
• c. Specialty filters and search options
•      e.g. people’s names, depth, anchor, java
  applets, domain, host, image, link, title, url,
  file/media types, business document types

• UNDERSTAND your engines
• http://www.beaucoup.com
         New developments
• Hyperbolic browsers :
  http:www.webbrain.com
• Semantic web
• Vertical search engines
• http://searchmil.com/
• http://www.corrosionsource.com/
• http://www.google.com/linux
         New developments
• OpenSource in an academic university
  library
• Samba : http://www.samba.org
• Apache Web Server :
  http://www.apache.org
• Majordomo :
  http://www.greatcircle.com/majordomo
    CLASSIC VERSUS FREE
    INTERNET SEARCHING
• What can you expect or what should you
  expect ?
• Example of a citation (web of knowledge
  and NEC)
• Example of a topic : cellular neural
  networks and visual computing
               B. Exercises
•   Panamarenko in the title-field
•   Laplace in url-field
•   Searchenginewatch.com in host-field
•   Osama Bin Laden in image-field