Search Engine Showdown
home feature chart reviews statistics learn directorires search

AltaVista Inconsistencies

AltaVista is one of the largest and most popular general Web search engines. However, it does not always behave as advertised nor deliver the results expected. This page aims to document both ongoing and short-lived inconsistent behavior on AltaVista. For more details on how AltaVista is supposed to work, see the AltaVista Review.

TIME OUTS: In order to provide results quickly, AltaVista may stop processing a search and only provide partial results. So the searcher can never know if all results have been retrieved. Repeating the exact same search several times in a row may find additional results. This problem and the inaccurate counting may also influence many of the other inconsistencies reported here.
Status: Ongoing

COUNTING: AltaVista can't count accurately. The numbers present are often estimates and can change greatly, especially with complex search statements that use multiple fields. For example, on 5/20/99 title:highsmith and not url:highsmith found about 60 hits. Then, on trying to go to the 6th page of results, AltaVista changed to "about 154 results." Trying the 15th page made AV change its mind to "about 77 results."
Status: Ongoing

LIMITS: Adding a language limit can cause some strange results including making the total number of results increase. An example reported Nov. 16, 1999 on Web4Lib demonstrates this. This problem only seems to occur on the simple search, not the advanced:
   +diabetes +domain:no found 95 (sites with hits)
   +diabetes +domain:no with Norwegian Language limit found 1498 (sites with hits)
   diabetes with Norwegian Language limit found 92 (sites with hits)
The second set should be smaller than the first, and the third set should be larger than the second. To avoid this kind of problem, use the advanced search.
Status: Ongoing

CASE SENSITIVITY: A search in lower case is supposed to find all occurrences of the term in lower, upper, and mixed case. However, sometimes it does not work like that. For example, a search on santa Fe found more than santa fe in the Simple search but worked properly in the advanced search. The Santa Fe problem turned up in a posting to the Web4Lib email list on May 10, 1999. It was fixed and functioning normally by May 15.
Status: Fixed

DIACRITICS: Similar to the case sensitivity inconsistency is the diacritic inconsistency. According to AltaVista's documentation, entering a search terms with diacritics will only find exact matches of the term with exactly those diacritics. Searching without diacritics will find the form with or without the diacritics. Thus, éléphant should only find the French form of the word, while elephant should find both French and English. However, if this always worked accurately +éléphant -elephant should find zero. As of April 1999, it found hundreds of hits. Test here to see how it does today.
Status: Ongoing

FIELD SEARCHING:

url: field Starting in March 1999, the url: field search reacted strangely. See also the What Did AltaVista Just Do? thread on Search Engine Forums for reports on this and a variety of other unusual behavior. The url: field is supposed to find the character string anywhere in the URL; however, during the problem time, it would not correctly find the search string within the domain name. The host: field search worked fine.
Status: Fixed as of May 1999

title: field The title: field search has been reported to not always find all records that have the search terms in the title. More on this problem will be posted after further analysis.
Status: Ongoing?

A Notess.com Web Site
©1999-2007 by Greg R. Notess, all rights reserved
Search Engine Showdown
Greg's Writings
Greg's Presentations