Corporate System
Third on the list appear ready-made solutions based on search technologies. They
are intended for serious companies and corporations with a truly gigantic
databases and under different instruments filled nick information systems. In
principle, the technologies themselves can be used for "home" needs. For example,
a programmer who works out of the office will use search technology for
scattered across the hard drive of his computer, the source of various programs.
But this particular. The main application of these technologies are still in
when required to carry out a quick and accurate search in large data files, and
for working with various information sources. The scheme of such search engines
is quite simple (although it is, of course, lots of unique methods of indexing
and query processing): phrase search. Given all stemming, synonyms and other
things. What leads us again to the problem of human resource: when using this
technology (which, for example, Hummingbird Line "program" or "Verity"), the
user must start to develop a tentative phrases that will be the search criterion,
and will, the idea to meet in appropriate documents. But far from the fact that
the user will be able to independently choose or you want to recall the words
and again, is not really sure that the subsequent search will be satisfactory.
Another key point is the speed of processing the request. Of course, when you
use a phrase to search an entire document instead of one of two words, the
accuracy of search increases in the order. But to date, just such an opportunity
has not been used because of the very high-capacity process. The fact is that
the search for a number of words or short phrases will not give us an accurate
resemblance discovered documents. A search for a phrase in the entire length of
the document in existing systems takes too much time and very demanding of
computational resources. There are conditional example. When refining search
request one word not much difference in what speeds will otrabotan this request:
0,1 seconds, or 0,001 seconds, as a user and not notice the difference. If you
take the same average size document which contains about 2000 unique words, the
searches based on morphology (word of) and thesaurus (synonyms), and the
withdrawal of the list of relevant documents found in the case with the search
by key words will take several dozens of minutes (which is unacceptable to the
user). |