The classification of retrieval systems is more convenient for building all on the basis of on the tax and processing of the information given to the users, - or else how many is automated in them, who types(collects) a database, in which the search is made: the people or computers.
2. Tools of search
Conditionally tools of search are subdivided into search means of a help type (directories) and retrieval systems in the pure state (search engines).
2.1 Thematic catalogues
The search tools of the first type name subject, or thematic catalogues more often. The company owning such catalogue, continuously conducts huge work, investigating, describing, cataloguing and displaying on half-glasses contents of WWW-servers and other network resources scattered worldwide. Result her(it) titanium of efforts is the constantly updated hierarchical catalogue, at the top level the most general(common) categories, such as "business" are assembled, ""science", ""art" etc., and the elements of the lowermost level represent the links to separate WWW-pages and server together with the brief description of their contents.
Guarantees that such catalogue really covers all contents WWW, nobody will give, however possible(probable) not the completeness and selection of materials is with interest expiated there, that for the present not under force to any computer - intelligence of selection.
The subject catalogues give also opportunity of keyword search. However search this occurs not in contents of WWW-servers, and in their brief descriptions stored(kept) in the catalogue.
The subject catalogues Internet can be counted literally on fingers, as their creation and support require(demand) huge expenses. To most known concern Yahoo, WWW Virtual Library, Galaxy and some other.
Is most popular in the population Internet the catalogue Yahoo. On the first page Yahoo, located to the address http://www.yahoo.com, you receive access to two basic method of work with the catalogue - keyword search and hierarchical tree of sections.
Having begun descent(release) on sections of the catalogue, you will see, that each section contains the precisely same field for input of keywords and button Search, starting search.
Each section can include as transfer included in him(it) and actually links to pages concerning at once all section, with their brief descriptions.
Instead of travel you can at once get in the necessary place of the catalogue Yahoo with the help of search. Having entered one or several keywords divided(shared) by blanks, in a line of search and having pressed the button Search, you receive the list all in Yahoo, that comprises the specified keywords. This list will be divided(shared) into two parts - "categories" and "sites".
If total of the links returned as a result of search, exceeds 25, the list of the links will be broken on some parts.
But not seldom happens so, that the list, given out by the machine, is very great and to see (overlook) it(him) simply not really. An output(exit) from this situation can become stricter selection of the information brought in the catalogue. One of most known such systems - catalogue Magellan to the address: http: // www.mckinley.com
This database contains the items of information on 80 thousand WWW-pages - that very much not much in comparison with those in millions, which exist in a network. However if Yahoo as the description of a resource uses one - two lines of the text, the employees of system Magellan on some of pages brought in their database, write the small reviews, and also estimate quality of these information resources on a five-mark scale. Till past bases of the reviews, Magellan owns as an own automatic index, for search in which it is necessary to throw the switch under a field of input in a rule(situation) entire database.
As a rule search represents one or several keywords divided(shared) by blanks.
The service, similar by the principles, of firm Point (http: // www.pointcom.com) in general basic emphasis does(makes) not on search, and on work with the thematic catalogue.
The service Point is known in a network for that its(her) employees are constantly engaged оцениванием of network resources and conduct the lists of those sites, which they consider(count) belong to " to the best five percents(interests) WWW ".
Firm Point Conducts a shared database all " of five-percentage WWW-pages, where about everyone it is possible to read the detailed license.
The most old subject catalogue WWW is the catalogue Virtual Library:
http: // www.w3.org/hypertext/DataSources/bySubj ect/Overview.html
This system full enough covers a scientific layer WWW - servers of universities, laboratories and educational institutions.
Russia-On-Line Subject Guide.
For the users in our country the certain interest can represent the thematic catalogue Russia-On-Line Subject Guide, located to the address http://www.online.ru/rmain. This catalogue contains rather motley assembly of the links on foreign sources plus the thematic review of the Russian and Russian resources WWW.
2.2. Automatic indexes.
It is possible to approach to a problem of search of the information in Internet and on the other hand. There are programs in which have loaded some thousand well-known URL-addresses. Being is started on the computer with access to WWW, this program begins automatically to download from a network the documents on it URL, and from each new document she(it) takes all links, contained in it,(him,) and adds them in the base of addresses. As at the end all WWW the documents are connected among themselves, early or late such program will bypass all Internet.
Certainly, the program can not understand as or classify that she(it) sees in a network. The programs of such type refer to as robots. They are limited to the tax of the statistical information and construction indexes in the texts of the documents. The database, collected by the robot, - index - stores (keeps) in it, simply speaking, item of information on that in what WWW-documents to contain those or other words.
Such the automatically collected index also underlies retrieval systems of the second sort, which frequently and name - automatic indexes.
The automatic index consists of three parts: the program - robot collected by this robot of a database and the interface for search in this base, with which the user works. All these components quite can function without intervention of the man.
As any classification of materials in such systems are absent, it is necessary to resort to them only then, when you precisely know keywords concerning that it is necessary, - we shall tell, a surname of the man or it is enough some of rare terms from the appropriate area. If to set search on the a little widespread words, you will have not enough life to bypass all URL-addresses, received as a result of search, - for example, the index of system Alta Vista contains of 11 billions words taken from 30 millions of WWW-pages.
An automatic index of WWW-pages exists much: WebCrawler, Lycos, Excite, Inktomi, Open Text and others. Some of them (for example, Lycos) represent more or less successful synthesis of the subject catalogue and automatic index.
Its(her) address http://altavista.digital.com. This system has appeared in December 1995. She (it) one of largest on volume of indexes from all such retrieval systems both most powerful and floppy rules of construction of searches. Alta Vista understands two different languages of searches rather strongly distinguished from each other. On the first page Alta Vista you see the form for simple search (Simple Search), and the panel of heading at the top of page contains the button Advanced Search, having pressed which, you receive the form for complication of search.
Except for WWW-pages, Alta Vista conducts a separate index for clauses from more than 14000 conferences Usenet (including hierarchy of groups relcom. *).
Search Alta Vista: that Alta Vista worked on group of words, only when they cost(stand) beside, it is necessary to conclude this group in inverted commas. If it is necessary to exclude from result all documents containing a certain word, it is necessary to attribute this word with is familiar "minus".
The word without any mark works in search precisely the same as also it with is familiar "plus".
As against Yahoo, by default Alta Vista searches of entry of the whole words. The ordered terms should stand in the document separately, instead of to be a part of other chains of symbols. If you need to find of all entry of a word, even when it is included into structure of other words, use a symbol *. The asterisk can stand only at the end of a word, and prevent giving many (too much) of results, Alta Vista requires(demands), that the word which is coming to an end on *, should consist not less than of 3 letters. Moreover, a symbol * allows to find not any termination (ending) of a word, but only not exceeding length of five symbols and not containing of capital letters or figures.