[ Contents ] [ Page Index ] [ WWW-Search ] [ Comments ]
VAN STARREX Website
NOTE: In the years since this web-site was extablished www.google.com has established itself as the leading search-engine. I have a Google search toolbar built into my web-browser and will soon add a dedicated search engine here when this page is next updated.
This page features all of the major WWW search-engines, together with instructions as to their features, use and search syntax. Leaving aside the specialised search engines that appear elsewhere on this site, this is really the first place to go if you want to find something out there in cyberspace. Links and other search-engines can be found on the 'Internet Search Engines' page, intsrch.html.
Searches using these search engines are done in one of three ways (in order of increasing complexity):
Getting to know the special features of any search engine that you use also helps cut down unproductive time spent sifting through less successful or simply over-productive search results.
First engine to handle natural language queries, this claims to have the largest Web index (8 billion words, 30 million Web pages). Also provides full-text index of more than 13,000 newsgroups and can find which sites are linked to a particular URL. Search results can be refined, though old and broken links can often be a problem.
Provides simple and advanced searches. Simple search allows "" (phrases/quotes), + (AND), - (NOT), * (word-fragment expansion - e.g. fire* finds fireman) and meta words: title:, anchor:, text:, applet:, object:, link:, image:, url:, host:. Also news article syntax: domain:, subject:, summary:, keyword:, newsgroup:. Advanced search does all this and provides for NEAR and OR Boolean and ranking of searches e.g. (fast AND car) NEAR race. Both types allow for refining of searches through requiring or excluding commonly found terms in results.
Has a unique AI (artificial intelligence) based navigation technology, allowing for 'fuzzier' searches - based on a statistical model of words appearing near each other in found documents. Full text of over 12 million pages and updated weekly, also usenet and newsgroup databases. Excite's journalists offer reviews of over 150,000 sites.
Searches are case-sensitive and Boolean words must be in capitals: +, -, "", AND, OR. Word-fragment expansion may take place unless a full stop comes at the end of a word, or will be done if $ is placed at end. Meta-tags in pages are ignored, as AI also used in link descriptions. Also has a 'more sites like this' feature.
Licensed to Wired/HotWired, this is probably the most powerful search engine available - indexing 10 million pages a day. Makes extensive use of cookie technology to store user demographics. Looks for keywords, first in meta tags - if none, in first few hundred characters. Can search web, usenet or both.
Searches allow any words, "", & (AND), | (OR), ! (NOT) and complex parentheses e.g. ((a OR b) NOT c) AND (d OR e). Allows metawords in keyword:value syntax - title:, domain: (URL), depth: (of pages), feature: (e.g. applet, audio, frame, image, vrml), linkdomain: (links to URL), linkext: (eg linkext:gif), newsgroup:, title:. Advanced searches allow for date, location, media type and page type. Date features work as either before or after day/month/year, or within number/unit e.g. within (last) 3/days. Location can be limited to type of site (e.g. .edu, .com) or place (GeoPlace).
Judged the most user-friendly search engine, and thought capable of providing the most relevant results. Has expanded in concept over the years, to provide a more complete information retrieval service - with email addresses, stock quotes, company profiles, and more. Is also starting to look more like Yahoo!, in terms of providing hand-selected site entries.
Accepts proper names with capitals (though regards any two capitalised words together as a proper name, unless separated by commas), otherwise handles capitalised words in quite a strange and complex way. Don't insert spaces between Boolean signs and operative words (e.g. +term). Features automatic expansion of words (so word-fragment* not necessary), together with metas: link:domain (links to URL), site:domain (sites within domain), url:x (pages with x in URL), title:x (pages with x in title), and pipes (like AND) e.g. space|asteroids.
Released July, 1994, this now has a database of several million entries - more than any other search engine. It can find text, graphics, sounds, and videos.
Reformats the output from several simultaneous major search engines into the one page - but also features their individual ads. Searches allow for + and - only (i.e. no more sophisticated searching syntax is allowed).
Searches 20 major search engines at once - including 411 (address search), FTPSearch and DejaNews. Faster but less reliable than MetaCrawler, and does not reformat the various search results.
The oldest browsable, searchable web directory i.e. sites are divided into categories. Also includes gopher and telnet searches. AND and OR are used like in AltaVista. NOT is used to divide classifications e.g. science NOT fiction.
First full-text crawler (1994), though it doesn't spider entire sites at a time. Unique and very powerful, though cosmetically now resembles Excite since corporate merging. Also features a list of the 25 most visited sites on the Web.
One of the first Internet guides, Yahoo! started as a collection of links, hand-picked and organised into directories. This is still the main way that it functions, though it ties to AltaVista for WWW-searches, and also features news, weather, sports and stock information.
Information is stored according to supplier-provided site descriptions. Searches are case-insensitive, and allow for (in order of priority) +, -, t: (titles only), u: (within URL only), "", *.
Search Options | Help on Search | Advanced Search Syntax
[ Previous Page ] [ Top of Page ] [ Next Page ]
[ Contents ] [ Page Index ] [ WWW-Search ] [ Comments ]
© Alex Van Starrex