Search Systems


Benefits of search systems:

  • Useful when there is too much information for a user to easily browse.
  • Can help mesh together fragmented sites.
  • Act as a learning tool for the user.
  • Can help tame dynamism.
  • Provides a well-known navigation tool that users will often expect to be there.

Basic anatomy of a search system:

Choosing what to make searchable:

  • Indexing everything isn’t always the ideal solution.
  • In large sites it can be better to create “search zones” – pockets of more homogenous content (e.g. technical reports, staff directory).
  • Some pieces or “atoms” of content should be ignored – e.g. the legal disclaimer at the bottom of every page.
  • It can be worthwhile to make the good content of your site searchable on its own, as well as being part of the site-wide search.
    • You can also program your search engine to search this good content first, then the rest of the site’s content if no useful results were found.

Determining search zones:

  • Search zones: subsets of a website that have been indexed separately from the rest of the site’s content.
  • User is able to narrow their search by selecting one or more search zones, improving retrieval performance and providing more relevant results.
  • Search zones can be created in as many ways as it is possible to to physically segregate documents or logically tag them.
  • Organisation schemes can often help in determining search zones, meaning the following can be the basis for search zones:
    • Content type
    • Audience
    • Role
    • Subject / topic
    • Geography
    • Chronology
    • Author
    • Department / business unit
  • When retrieving results, destination pages (which display the actual information sought) should be included in the output, while navigation pages (pages that assist the user to get to destination pages, such as main pages, search pages and browsing assistance pages) should be removed to avoid clutter.

Search Algorithms:

  • Pattern-matching algorithms: Most retrieval algorithms use pattern matching; typically a search query is compared with an index of the full texts of an entire website’s documents, with the documents containing matching strings being added to the retrieval set.
  • Recall and precision: Some algorithms return numerous results of varying relevance, others return a few high quality results.
    • Recall = # relevant docs retrieved / # relevant docs in collection
    • Precision = # relevant docs retrieved / # total docs in collection

Query builders:

  • Query builders are tools that help improve the performance of search engines.
  • Examples include:
    • Spell-checkers
    • Phonetic tools
    • Natural language processing tools
    • Controlled vocabularies and thesauri

Presenting results:

  • There are 2 main issues to consider once results have been assembled.
    • Which content components to display for each retrieved document.
    • How to list or group the results – e.g., sorting by alphabetical order, sorting by chronology, ranking by relevance, user’s or expert’s ratings, popularity, pay-for-placement.

Designing the search interface:

  • Things which affect the type of design required:
    • Level of searching expertise and motivation.
    • Type of information need.
    • Type of information being searched.
    • Amount of information being searched.
  • The box – users will have lots of assumptions about the functionality of an organisation’s search box; these should drive the default settings on the simple search interface.
  • Repeat the search query in the box on the results page.
  • Explain where results come from: e.g., 10 scientific journals, 5 newspaper articles, 5000 technical support pages.
  • Explain what the user did, if results aren’t satisfactory.
    • Restate the query.
    • Describe what content was searched.
    • Describe any filters that might be in place (e.g., date ranges).
    • Show implicit boolean or other operators, such as AND.
    • Show current settings, such as the sort order.
    • Mention the number of results retrieved.
  • Look for opportunities to integrate searching with browsing.

When users get “stuck”:

  • If there are too many results:
    • Provide tips on how to narrow search results.
    • Allow users to search for items within their current result set.
  • If there are too few results:
    • Adopt a “no dead ends” policy with regard to zero searches, meaning users always have another option. Options can consist of:
      • A means of revising the search.
      • Search tips or other advice on how to improve the search.
      • A means of browsing (e.g., including the site’s navigation or sitemap).
      • A human contact if searching and browsing won’t work.

Chapter 1: Search as a system
The user’s view of the search system, search system components and metrics, SharePoint search system architecture

Chapter 2: The SharePoint search system
Overview of SharePoint search system features: content, policy, search functions, site navigation, application integration as well as what’s missing

Chapter 3: Planning & governance
Governance vs. planning, planning approaches, tasks, team, worksheets, workshops, and governance models

Chapter 4: Where to store content
SharePoint content storage options, SharePoint vs. external content repositories, usability considerations

Chapter 5: Basic setup
Administrative interfaces, basic setup tasks, setting up the search system (content sources, crawl rules and crawling content, planning, creating and exposing search scopes)

Chapter 6: Metadata setup
How metadata is stored in SharePoint, how to map crawled properties to managed properties, the new book function, how to use the Term Store Management Tool

Chapter 7: Customizing SharePoint search
Search page components, relevancy ranking, keywords and Best Bets, how to create new search pages

Chapter 8: Customizing search results
How to change the metadata in the search item description, how to use SharePoint Designer to change the results page look and feel, how to add a custom Web part

Chapter 9: People search
People search vs. content search, sources of people data, how to customize people search

Chapter 10: Content types
How to create and use custom columns, content types, and document templates, how to change document metadata

Chapter 11: Customizing navigation
The role of navigation in the search system, navigation components, how to customize navigation

Chapter 12: Metadata integration
Metadata management options, how to link to an external metadata repository, a custom Best Bets Web part, third-party products for metadata integration, morphological processing

Chapter 13: Special topics
“Zero term” search with the content query Web part, audience targeting, searching images, integration with desktop applications, multi-lingual search

Chapter 14: Improving the search system
Usage reports and surveys, observation and analysis, collaboration profiling


Leave a Reply

Your email address will not be published. Required fields are marked *