This page show you how to submit queries to Altavista.
A more complete help can be found here.


paris "petite galerie" louvre
Finds documents containing as many of these words and phrases as possible, ranked so that documents with the most matches are presented first.
A phrase is any string of adjacent words. The preferred way to link words into a phrase is to use quotes.
Lower-case search will find matches of capitalized words also. For example, paris will find matches for paris, Paris, and PARIS.
Capital letters in a search will force an exact case match on the entire word. For example, submitting a query for parIS will search only for matches of parIS. (Don't be surprised if there are none.)
+noir +film -"pinot noir"
Matches may be required, or prohibited. Precede a required word or phrase with + and a prohibited one with -. This query finds documents containing film and noir, but not containing pinot noir.
antique;pump;organ
Punctuation glues words into a phrase, just as quotes do. Punctuation is treated as white space, so this example is equivalent to "antique pump organ" (that is, three words enclosed in quotes).
quilt*
This query matches pages that contain at least one word such as quilt, quilts, quilting, quilted, quilter etc. Hint: The *-notation is also useful for searching for variant spellings. For example, alumi*m will find matches for both aluminum and the British English aluminium. More about its use


Examples of Simple Queries

To find the documents most relevant to what you need, construct your query as precisely as you can. AltaVista ranks the documents found so the ones matching the most words and phrases in the query are listed first. Even so, you might not find exactly what you want at the head of the list if your search is too general.

For example, suppose you wanted information about the languages of American Indians but you did not know any specific language to search for. You might start with the following query: american indian language. (The word-count numbers quoted here are not updated as new pages are indexed. They serve as an example only.)

american indian language
result:
word count: indian 395185, language 2048030, american 2654433. 100000 documents found containing as many of these words as possible, in both upper and lower case.
observation:
This search is much too broad. Of the first ten documents found, the first few appear relevant, but the rest are documents about languages in the Asian subcontinent.
strategy:
Make clear how you want the query to be parsed. In other words, link american and indian together as a phrase. Include the plural of language in the search also by using the *-notation.

"american indian" language*
result:
word count: american indian 30000, language* 2050463. 20000 documents found.
observation:
The documents found are now relevant to information about American Indian languages, enabling you to refine your search further. For example, suppose you want to know more about the ojibwe language that was mentioned in one of the documents found by this query.
strategy:
Require that the word ojibwe and its variants ojibway and ojibwa be included in your next search. Since this is an American Indian word, you could now omit american indian from the search.

language* +ojibw*
result:
word count: ojibw* 3625, language* 2050463. 1000 documents found.
observation:
Bingo!

Ranking Simple Queries

For Simple Queries, AltaVista will rank the results based on a scoring algorithm; documents with a higher score appear at the head of the ranking list. A document has a higher score if the following hold:

You are therefore likely to find what you want close to the head of the resulting list of matches.


Constraining searches

It is possible to restrict searches to certain portions of documents by using the following syntax. The keyword (link, title, image,...) should be in lower-case, and immediately followed by a colon.

Constraining searches in Web pages:

anchor:click-here
Matches pages with the phrase click here in the text of a hyperlink.
applet:NervousText
Matches pages containing the name of the Java applet class found in an applet tag; in this case, NervousText.
host:digital.com
Matches pages with the phrase digital.com in the host name of the Web server.
image:comet.jpg
Matches pages with comet.jpg in an image tag.
link:thomas.gov
Matches pages that contain at least one link to a page with thomas.gov in its URL.
text:algol68
Matches pages that contain the word algol68 in any part of the visible text of a page. (ie, the word is not in a link or an image, for example.)
title:"The Wall Street Journal"
Matches pages with the phrase The Wall Street Journal in the title.
url:home.html
Matches pages with the words home and html together in the page's URL. Equivalent to url:"home html".

Constraining searches in Usenet news articles:

from:napoleon@elba.com
Matches news articles with the words napoleon@elba.com in the From: field.
subject:"for sale"
Matches news articles with the phrase for sale in the Subject: field.
You can combine this with a word or phrase. For example, subject:"for sale" "victorian chamber pots".
newsgroups:rec.humor
Matches news articles posted (or crossposted) in news groups with rec.humor in the name.
summary:invest*
Matches news articles with the word invest, investment, investiture, etc., in the summary.
keywords:NASA
Matches news articles with the word NASA in all caps in the keyword list.

More about Words, Phrases, Capitalization, Accents, and the *-Notation

Words

AltaVista treats every page on the Web and every article of Usenet news as a sequence of words. A word in this context means any string of letters and digits delimited either by punctuation and other non-alphabetic characters (for example, &, %, $, /, #, _, ~), or by white space (spaces, tabs, line ends, start of document, end of document). To be a word, a string of alphanumerics does not have to be spelled correctly or be found in any dictionary. All that is required is that someone typed it as a single word in a Web page or Usenet news article. Thus, the following are words if they appear delimited in a document: HAL5000, Gorbachevnik, 602e21, www, http, EasierSaidThanDone, etc. The following are all considered to be two words because the internal punctuation separates them: don't, digital.com, x-y, AT&T, 3.14159, U.S., All'sFairInLoveAndWar.

Only the words in a document are significant to AltaVista. AltaVista does not index punctuation or white space, so you can use AltaVista to look only for words and phrases, not punctuation.

Phrases

A phrase is a string of words that are adjacent in a document, although they may be separated by any amount of white space or punctuation. They do not have to be grammatical in any human language--they just have to occur in a document as an adjacent sequence of words. Some examples:

Since the punctuation and white space are insignificant to AltaVista (except that they delimit words), the phrases above are indistinguishable from the following variants:

There are two conventions for typing a phrase in a query. The best way, leading to the least ambiguity, is to type the phrase as "a sequence of words separated by spaces and surrounded by double quotes". However, as an alternative, you may type the words of the phrase with punctuation (and no white space) between each pair of words. For example, these are all equivalent as queries:

The first is the one we generally recommend. Be aware that the punctuation characters & | ! and ~ have meaning in Advanced queries, and * indicates the *-notation used in both Simple and Advanced queries.

Capitalization

Capital letters are considered distinct from lower-case letters. When a word is found in a Web page or a news article, its case is preserved when it is stored in the index.

When you enter a word in a query, therefore, it is always safe, and generally recommended, to type it all in lower-case, because lower-case letters indicate a case-insensitive match. If you type any capital letters, you force an exact case match on the entire word.

Thus, the word turkey in a query will match any of turkey, Turkey, tUrKeY or TURKEY occurring in a document. But the capitalized word Turkey in a query will match only Turkey in the document, and not any of the other capitalization variants.

Accents

Accents are treated in the same way as capitalization. An accented word used in a query forces an exact match on the entire word. For example, if you use éléphant in a query, you will match only the French spelling for the pachyderm. However, if you do not care to enter accents in the search window (something which is browser, platform, and keyboard-dependent), you can always safely omit the accents, thereby matching both the French and English spellings.

The *-notation

To search for occurrences of any of a group of words with a similar pattern, AltaVista provides the *-notation. For example, you might want to search for matches of sing, singer, singers, singing. In this case, place the *-notation at the end of the word whose inflections you want to include in the search: sing*. But, a word of warning. AltaVista will also match words lexically unrelated to your query word. So the query sing* will also find matches for singe, single, singular, and for foreign words such as French singulier.

The *-notation cannot be used without restriction. To make such queries computationally feasible, AltaVista requires that the * be used only after at least three letters. The *-notation will match from zero up to five additional letters in lower-case only. Capital letters and digits will not therefore be matched.

The *-notation can sometimes be useful for finding variant spellings: for example, cantalo* will find matches for cantaloup, cantaloupe, cantalope, and their plurals. But take care how you construct the query word. For example, if you want to find matches for both color and colour, a query of the form col*r is not the most efficient. This query will also find matches for collector and atomic collider. In this case, it is more efficient to submit the query colo*r, which will find matches for both color and colour.

Finally, if your search using the *-notation finds too many matches, AltaVista will ignore the query. The query inte*, for example, produces the result,

Ignored  inte*: 4292323  
No documents match this query


The META tag: Controlling how your Web page is indexed by AltaVista

In the absence of any other information, AltaVista will index all words in your document (except for comments), and will use the first few words of the document as a short abstract.

It is however possible for you to control how your page is indexed by using the META tag to specify both additional keywords to index, and a short description. Let's suppose your page contains:

<META  name="description" 
content="We specialize in grooming pink poodles.">
<META  name="keywords" content="pet grooming, Palo Alto, dog">
AltaVista will then do two things:
Pink Poodles Inc
We specialize in grooming pink poodles.
http://pink.poodle.org/ - size 3k - 29 Feb 96

Copyright © 1996 Digital Equipment Corporation. All rights reserved.