AltaVista

AltaVista was created by researchers at Digital Equipment Corporation (DEC) who were trying to showcase the company's hardware to make it easier to find files on the company's public network. Paul Flaherty came up with the idea, along with Louis Monier, who programmed the web crawler, and Michael Burrows who wrote the indexer. Ilene H. Lang was the first CEO of AltaVista; she was recruited by DEC to build its software business. At launch, AltaVista had two innovations that put it ahead of other search engines: It used a fast, multi-threaded web crawler, Scooter, that could cover many more World Wide Web pages than were believed to exist at the time, and it had an efficient backend database, TurboVista, running on advanced hardware. In 1997, it launched Babel Fish, a web-based machine translation application that translated text or webpages. In January 1998, Compaq acquired DEC for $9.6 billion, including $4.8 billion in cash and the remainder in stock, to acquire its server technology. As part of the transaction, Compaq acquired AltaVista. In February 1998, AltaVista added free webmail. In May 1998, it ended its deal with Yahoo to provide search results. In August 1998, Compaq acquired the altavista.com domain name for $3.3 million and redirected the site; it had previously used altavista.digital.com. Until January 1999, AltaVista featured only a search engine on a "utilitarian" web page with simple yellow and orange graphics. In February 1999, Compaq acquired Zip2 to enhance AltaVista. In June 1999, Compaq sold 83% of AltaVista to CMGI for $2.1 billion in stock and $220 million in cash. In July 1999, AltaVista began offering financial content. In August 1999, AltaVista began offering free dial-up Internet access. It ended the offering in December 2000. In October 1999, the site, as well as its price comparison site Shopping.com, was relaunched. It then offered real-time traffic pictures, news, financial information, and entertainment. It also increased staff from 80 to over 600 and launched a new tagline: "AltaVista: smart is beautiful". It aired television advertisements, including one with Pamela Anderson, as part of a $120 million marketing campaign. In November 1999, AltaVista acquired RagingBull.com. In December 1999, AltaVista filed for a $300 million initial public offering; it was cancelled due to the end of the dot-com bubble. In 2000, according to Media Metrix, AltaVista was used by 17.7% of Internet users while Google Search was used by only 7% of Internet users. In October 2000, Rod Schrock resigned as CEO. In February 2002, AltaVista shut its webmail service; at the time, it had 200,000 users. The accounts were transferred to Mail.com. In November 2002, the website was redesigned again to include a pared-down front page and more frequent updates of indexed links. In February 2003, Overture Services acquired AltaVista for $140 million. In July 2003, Yahoo acquired Overture for $1.63 billion. In June 2013, the domain name was redirected to Yahoo Search. ==How it worked==

There were two major search modes: simple querying and advanced querying. A "simple query" looked like `word1 word2 "phrase" -word3 +word4` which was interpreted as "(word1 OR word2 OR "phrase") AND NOT word3 AND word4". Words within double quotes were phrases: they must be adjacent in a document for the document to match the query. A "query term" was a word or a phrase. An "advanced query" was an explicit Boolean expression. In advanced query mode, `and`, `or`, and `not` were interpreted as Boolean operators rather than as search terms. Advanced queries may have also included `near`: the words on either side of `near` must be close -- but not necessarily adjacent. Both simple and advanced queries supported `host:xx.yy.zz` which queried only documents found on the hostname (web domain) `xx.yy.zz`. A pull-down menu allowed the user to restrict result pages only to pages in a particular language. In the advanced search, an input box allowed the user to restrict the results to pages last modified on a certain date, or within a range of dates. AltaVista returned URLs ranked by its internal "relevance function". Each page contained 10 URLs. The user may click on "3", for instance, to get to the 21st-30th most relevant URLs. This differed from some other search engines, where the user could jump to only the next 10 or previous 10 URLs. AltaVista logged user requests in a "query log" A request may consist of a new query or a new result screen for a previously submitted query. Each request included the following fields: Unix timestamp for the query; cookie (blank if the user had disabled cookies); query terms; result URLs; other user-specified query modifiers, such as a restriction on the result pages' language or date of last modification; metadata, such as whether the query was a simple or an advanced query, the browser the submitter was using, the IP address of the submitting host, etc. AltaVista collected session information to study querying behavior. A "session" was a cluster of queries in a short time by a single user. Queries with the same cookie were assumed to come from the same user. For those 4% of queries in which the user had disallowed cookies, a cluster of queries coming from the same web browser on the same domain IP was treated as a session. This method was a poor substitute for cookies, particularly on large ISPs such as AOL, where the number of users sharing a single IP address could be around 10,000. ==Further reading==