Overview of Google Search options

Summary of available information about the 5 google search options:

The free options

The Ajax Application Programming Interface (API)

  • Javascript tools that allow you to do neat stuff with Google search results; option to include without JavaScript using PHP
  • Possible to combine this with custom search to do lots of neat things (code example)
  • Allows customization of Cascading Style Sheets (CSS)
  • has a limit of 32 results over 4 pages (8 results/page; limit of 8 results for local/blog search filters); otherwise no query or document limit

Custom Search Engine

  • currently being used by the Library
  • allows you to combine domains to create a custom engine
  • can customize CSS & integrate into your web template
  • universities exempt from advertising requirement
  • has some stats reporting (??? BIll ?? Specifics ??? Has a popular queries list)
  • http://www.google.com/coop/docs/cse/cref.html[Linked CSE]] allows you to do some of the definitions on your own server; might solve some of our problems. (e.g. can automatically generate CSE's)
  • may be a limit to the number of results returned; haven't found much information see Google Custom Search Help
  • seems to require JavaScript as well

The old Public search ("old" search)

a.k.a Google Syndicated Search, University search

  • The old public search we're using in "old" search was quietly discontinued by Google sometime in 2006/2007. It still works but they are no longer actively promoting it. Custom Search seems to be a replacement for this service
  • In late-2006 the login for the admin interface was disabled due to security problems. It seems like they never put it back up again (although the custom templates do work)

The Google Simple Object Access Protocol (SOAP) API ("new" search)

  • this API was provided for a short time in 2006. Like the "old" search, the keys still work, google just isn't providing new keys anymore.

The Enterprise Options

Site Search

Features:

  • results in extensible markup language (XML); API allows you to access this
  • comprehensive indexing
  • public content only
  • can control the crawl (frequency, depth & breadth)
  • can adjust the results (date biasing, top results biasing)
  • synonyms, keywords, query expansion, subscribed links (like keywords)
  • customize display
  • some reporting features (not sure of details)
Mini appliance
  • Like enterprise appliance but limit of 300,000 documents

Enterprise search appliance

  • appliance (you buy the box)

Features:

  • results in XML
  • comprehensive indexing
  • can search protected content
  • public content only
  • can control the crawl (frequency, depth & breadth)
  • synonyms, keywords, query expansion, subscribed links (like keywords), "onebox" results
  • customize display
  • has some reporting features (good, not great; would still want to connet to analytics)
  • can do decentralized administration (keymatches, stylesheets etc.)

Sample university sites

-- MeganMcDermott - 16 June 2008