N.Y.Zaitseva, E.A.Teliatnikova ANALYSIS OF THE TERM FORMATION WAYS IN THE FRENCH LANGUAGE (WITHIN THE FRAMEWORK OF A SUBLANGUAGE OF THE TELECOMMUNICATION SOFTWARE MEANS)

E. Tsybina ADVANTAGES AND DISADVANTAGES OF USING THESAURUSES

Due to the data search system there appeared ideographical dictionaries that can make the task of data search (especially documental) easier. Such dictionaries are called Data Search Thesauruses. A thesaurus doesnt contain a list of words only. Its a rather complicated device. Each DST is supposed to retrieve information in a specific subject area. In comparison with defining dictionaries that describe meanings of words and notions, the thesaurus gives certain information about the language as a whole and the scientist is able to deduce word and notion meanings himself. The DST reflects the semantics of the language much better than terminological dictionaries. It also describes synonymy, homonymy and meta language semantics as a whole. If the language is rather difficult, the thesaurus helps to reduce polysemy.

Owing to a great deal of information, detection, exclusion and pertinance of terms necessiate the use of thesauruses. Thesauruses are also meant for detecting the meaning of the object under consideration, for synthesizing and analazing knowledge, for classifying it in a specific subject area, for decision making and extracting some heuristic prompts from them. The extention of knowledge about thesauruses gives a possibility to enlarge requests to the DST.

Like any object, property or relation the thesaurus has its own advantages and disadvantages. First of all, the thesaurus helps reduce polysemy in the natural language: here its quite possible to substitute one notion for another and it will cause no changes in the document relative to the data search. The thesaurus can also be a link that can organise the dialogue “human-computer” in the data search system.

Its quite evident that thesauruses are good at putting tasks of a specific subject area into practice. But, nevertheless, they are incapable of covering all the words of the natural language thats why it can be confirmed that the process of indexing can sometimes distort the contents of a document. Besides, the traditional thesaurus is rather static. Consequently, it doesnt reflect the dinamics of the language very well. The DS Language lacks the word control when entering the system. The working out of thesauruses is labour-intensive. So, it cant guarantee the high quality of data search.

The practice shows that the specialist fulfils all the prelimenary routine while working at the dictionary. During the coordinate indexing he uses certain rules that let him find out how the pertinance of a word depends on its position in a phrase. At the same time its obvious that the specialist has derived most of these rules from his practice work. Therefore, they remain implicit and the specialist cant formulate them. The explicity of these rules can become possible if we compare syntactical structures including these very words and the pertinance of these words. The pertinance can be defined by analyzing the specialists results, that is by comparing the cases when the specialist considers this or that word to be a key word. So, a number of key words for a certain document can be looked upon as a result of some psycholinguistic experiment. Analyzing this result one can formulate a list of rules that the specialist uses while recording the text of the document into the text of key words for this document. Since indexing is implicit, the usage of this rules is sure to be inconsequent. Hence, there are errors in indexing. Thats why its very urgent to find the way of formalizing the methodology of indexing.

So, the lexicography of thesauruses needs to be improved and this fact can help the specialist of a specific subject area solve certain tasks.

Purtov I., Tsybina E. THE DICTIONARY OF CONCORDANCES FOR THE APPLIED FIRST-ORDER PREDICATE LANGUAGE

The text of a technical task that the designer gets from the client, can contain some omissions and contradictions due to different reasons. Therefore, the performer can interpret this task in different ways. To take it into consideration, the performer should coordinate the text with the client. A special analysis of the text can become a means to detect all the omissions and contradictions. For instance, its formalization as the analysis with the translation into a suitable formal language. As an example of such means its recommended to use the translation of texts in the natural professional language to first-order descriptions.

The text of the task and all the demands included into it is a system of predicates. Some logical languages such as PROLOGUE correspond with these structures very well. Before translating the text in the natural language its necessary to detect the bearer of predicativity. Simple sentences are the main bearers. However, some turns of speech and other syntactical structures can take this part as well. In the research only predicativity on the level of sentences is used.

Not only Russian texts can be translated into the language of predicates. The English language is taken as the second one for translation because today it is the most widely spread language in international systems of data exchange. Besides, the quality of formalizatin can be better if we do the translation into 2 versions of Prologue-based description in both English and Russian languages.

For different natural languages there exist classifications of sentences according to semantical and structural properties. During the work there was made a table of typical schemes of Russian and English languages and their description in the language of predicates. A modern English classification was thoroughly studied where exists a division of sentences into complex and compound. The same goes for the Russian language. But in English ascentive sentences dont form a separate group as in Russian

There are 19 schemes of English complex sentences: 4 of them are compound and 15 - complex. There are also 2 constructions in the table which are neither compound nor complex but they are translated into Russian as complex.

The specialist translated the sentences of these types into Russian trying not to deviate from the meaning of the sentence and got the phrases that refer to 30 Russian schemes. In reality, in English grammar there are more types of English constructions. Such discrepancy is due to the fact that in Russian there exist more subtypes of sentences than in English. Such situations as, for example, when different relative subtypes are described in the Prologue language in different ways are also taken into consideration.

There are 9 ways of translation into the first-order language in the table that are reduced to 3 ones:

• 
  forming the rules of logical deduction
  
• 
  addition to the database of the program
  
• 
  forming of constructions that simulate predicates of the second order