Email: alexius87@rambler.ru
CASE Software has high enough urgency presently. Its application has enough considerable quantity of fields of activity and levels proffesionalism. In master's work CASE SW will be projected in economic area of work with the client. Work means creation CASE for selection the plan of development of the enterprise fo the investment with the minimum expenses. The result of program will be the model choosed by engine with the parametres, requested by the client. The client enters the necessary data, the system chooses approaching under this template business plans and offers them on consideration to the user. Main duty in this work os addition (completion, updating) algorithm of work with technical projects and ready variants of the decision.
There are some variants of algorithms of work with representation of texts in the form of a tree:
The algorithm, being able to make plural operations over texts, carries out association of texts in one [2]. Technical projects are also being combined in one tree. Then by detour of a tree of prototypes on all or knots the target sequence of parametres will be under construction. Usual algorithms work by a principle of dialogue with the incompetent user, which does not foreknow anything about a subject domain and cannot specify data independently wished for it, but can answer questions . In this case pasting goes without the analysis, a continuous method. In the given work the modified algorithm of processing of entrance texts will be considered. When ES contains base grammar for the analysis of texts and the user is competent for defining initial restrictive parametres. Distinctions of algorithms on the basic variants of entrance data will be considered below.
In this case the system does not know anything about a character set and keywords of texts and association will go "blindly". Usual association with transformation in and-or tree goes by a principle of comparison to a finding of the first differing element. For example, we will consider pasting of two texts of figures representing to the simple list. Association is made as is shown in drawing 1. It will be as a result generated i-or a tree as is shown in drawing 2. At a conclusion of target sequence on a tree the system will be compelled to ask what parametres are nessessery for user. For example at detour from left to right the system in a point «B» already having constructed the list {3,4} will ask the user what there should be a third parametre and depending on the answer, on an exit there will be a necessary sequence of parametres. This algorithm is not capable of training, that is for it the essence of the met symbol is not important, it is simply compared under number to a corresponding element of other text. Such in the image in the presence of often repeating elements the bulky tree as all elements will be included in it so much time how many they have met in all texts.
In this case on the same example the user can set in advance the parametres necessary to him. For example: one of parametres A equal 4 and parametre B equal 6. Then the system will simply unite sets and at tree detour will construct correctly target sequence of parametres. Pasting by a continuous method also concerns lacks of the given variant, that is without sorting and a filtration, including all repetitions that involves increase in quantity of the memory occupied with a tree.
If the user sets initial data about syntax, a base set of terms, synterms which meet in texts the system can distinguish texts on key data, that will simplify a problem.
Let's consider an elementary example - texts represent some code of the program. As is known in a code many repeating keywords. And at grammar use it will be possible at a finding defined synterm to finish simply in the list of repetitions of it synterm in texts the identifier of a prototype from which it is taken data synterm and a place where it has met (a line or consecutive number). That is the algorithm will be capable to assort that for synterm has arrived on an input. A stage of association of all texts in one and-or tree with marking dependences and repetitions we will name training as the system definitively creates full representation about target area of the system. Then the target sequence taking into account repetitions and dependences will be under construction and it will be easier to cut incorrect variants. For example if has met synterm which is present only at first third of five texts the second, fourth and fifth texts will be rejected and will not be considered more.
Working with the competent user
If the user is qualified enough, it sets completely all grammar, that is all possible terms, synterms, meeting in texts and dependence, where and when they meet. In that case the grade level is passed, as already all grammar of target area of the system is already known, and the conclusion of gentle sequence proceeding from the technical project begins.
Among the main tasks it is possible to list
As of date of a writing of the abstract, from published in a worldnet internet works it is possible to draw a conclusion, that only chair PMI is engaged in idea realisation продукций over грамматиками in the name of the senior lecturer of chair Grigoriev A.V.as of people working in this sphere of activity very little scientific value of work омжно to estimate as powerful.
The problem is actual enough for today. At the due approach and expediency of the decision, can carry both economic continuation, and scientific as the field of knowledge is still poorly studied also workings out in this sphere a little enough.