Automatic Conversion of Between Programs of Different Structures

7

[(]

An Study on Algorithm to Realize

Automatic Conversion of between Programs of Different Structures

Katsumi Yasuda, and Hamid Fujita

Abstract--AnIn this work a development procedure has been reached for implementing a new algorithms for automatic conversion of source programs with different structures is realized. In a conventional program, a distinctive correspondence between its structure and its user requirements can not be made. Precisely speaking, a concept of control logic is indispensable, which takes about 80% of a program. Such control logic is created by software development engineers based on the interpretation and engineering of the user requirements.This control logic is made by someone specialized in software development based on the user requirements. The control logic and the user requirements are un-recognizedably mixed up.mess. Such mixture betweenof the user requirement (bythe user) and engineering oriented requirements created (by the system engineers) for system design make the control logic specified by the system engineer more dominated parameter for the system performance. From our experience, this type of logic is the major problem for maintenance and system evolution,. due to that, tThe resulted requirements in general becomes difficult to be realizable due tobecause overloaded control logic specified by the engineering aspect of the software engineer are additional constraints. We call such a complicated situation where system requirements have on having complicated logic on system requirement, a mess. In order to make an algorithm for automatic conversion between programs, such a mess needs to be analyzed and interpreted, which is not possible by conventional impossible by syntactic and linguistic analysis. One possible way to create such an an algorithm is usually, semantic analysis whereas semantic analysis is done only by people not a computer. This may create difficulty for practical efficient user realization. However, we find such conventional realization is still very narrow and non-general for other type of applications programs. From our investigation we realize that Ffor automatic conversion, we need a method replacing the semantic analysis to make such conversion more generally efficient and respecting the actual user requirements.. In this work, weWe will discuss how such an algorithm can be created. The basis of the discussion is a program structure called Lyee[1]. The Lyee program structure is unique, in that the control logic does not appear in the development stage, differently from conventional software development. need to on purpose necessary (as in conventional software development) to ., Consequently, i.e., the control logic does not appear in the source programs depending on the user requirements. Our approach makes the user requirement as the major player in system design. The control logic is reflected and imbedded into Lyee universal structure implicitly.

Generally, usual application and user requirements are entangled with control logic at the development stage. Therefore, the resulted programs has 80% of control logic which make the program behavior predication complicated and difficult to analyze for possible evolution. Such complexity of having entangled control logic makes it uneasy to act on such programs for evolution purposes. However, through our studies, we realize that it is possible to realize such conversion algorithms between programs when those programs are converted into Lyee structured typed programs. This is possible because, Lyee typed structured programs does not have control logic and independent on programming languages.

Index Terms--—program structure, control logic, user requirementAbout four key words or phrases in alphabetical order, separated by commas (see "Subject Categories" in a recent January issue of IEEE Transactions on Magnetics, automatic conversion

I.  INTRODUCTION

CI

ONnventiolal order to realizeprograms are to replace user requirements into functions, (i.e., functions are system components). The functions are considered to be the source of meaningHowever, whereas Mmeaning is considered to exist only in in human mind. Functions are considered andto be expressed in our behavior, whose cause is brought about instantiated which originates in our mind. Since user requirements include both meaning and functions, isn’t it natural to consider that requirements cannot be expressed only with functions? Control logic of conventional programs possesses this contradiction. TheRealization of user requirement realization by the conventional development procedure (data oriented and structural oriented system architecture), creates that such logic whichthat represents the meaning of thate requirement, but the meaning is interpreted through a complicated structure mix up.

On the other hand, characteristics of characteristics of the structure of Lyee programs isis deduced from the Lyee theory so that the meaning of user requirements is to be is to be realized. In Lyee, meaning is replaced with ‘intention’ as a meta-model. The reason why intention instead of meaning is used is that intention is easier than meaning to deal with in the meta-model [1].[2] Lyee program structure has been based on such concept. That is the main structure of Lyee program based on a universal structure that we called as Predicate Structure [2]. The Predicate Structure which realizes an intention (i.e., user requirement) without the inherent needs of the logic needed by the engineering neededfor realizing to realize thosee requirement. Thatis engineering part (i.e., control logic) would be realized by the recursive structure of the pPredicate sStructure of Lyee. Therefore, thosee logic will be created at run-time and not readable by thean implementation program, but it is imbedded into the Lyee structure meta-model. This unique realization is the main motivation for our algorithm conversion approach, basinged on the Lyee methodology[2]. For more details on Lyee please referrer to the link (http://www.lyee.co.jp). user's defined requirement, conventional methods use engineering models to interpret the meaning of the requirement and write specifications on it, which are considered designing.

A. 1. Conceptual relation between Lyee program and conventional program

See Fig. 1.

An algorithm to convert Lyee programs into conventional programs has beenis realized, and is not presented which is not discussed in this paper. due to lack of space..

By

By analyzing a conventional program, a skeleton of the Lyee program structure is realized. The analysis is consequently,done using a program that can be made based on the concept of the Lyee program structure. Such resultedThis This program for analysis is one of the elements that compose the automatic conversion algorithm. The TheOurskeleton here means a synchronous state of a program to be converted using in Lyee methodology, whereas it means control logic in conventional programs. The synchronous state is a structure that realizes a memory in the Lyee theory [3].[3]

In the Lyee theory, existence in the natural world i is considered to be defined and presented inby the memory. For example, a sentence (i.e., requirements) represented written in a natural language is defined byin the memory. The memory is not necessarily that of humans. The memory is defined and instantiated by intention. GenerallyT, the aim of the Lyee theory is to grasp the intention. When the intention is grasped, the memory is realized. TSo the state onof realizing that realizes thera reelation between the intention and memory is called the synchronous state. The synchronous state is modeled and representedrealized made into a model called Three-dimension-like Space Model (TDM). The role of TDM in automatic conversion will be discussed in the next chapter. TDM is an element composing the algorithm for automatic conversion. There are two cases in which a correspondence between the conventional program to be converted and TDM is established:

1) One to N (N is a plural number)

2) N to N

The factor to determine either 1) or 2) above is Logical Behavior of items belonging to a logical record on a memory. The logical record is called and the Defined in Lyee framework, which is, for example, a set of items appearing on screens, files, memory tables, or communication messages. The items are called words, which are nouns expressed in the user requirements.

The concept of lLogical Bbehavior (LB) is defined by the Lyee framework theory as thea correspondence between a memory of a word and its uUnitized iInstructions Group (UIG). There are two cases of the correspondence.

1) One to one

2) One to N (N is a plural number)

The one-to-one correspondence can be replaced by Signification Vectors of L3 and L4, while one-to-N correspondence can be replaced by Signification Vectors of L2 and Input Vector. The analysis mentioned above is done by the program to analyze conventional programs. The algorithm of conversion is given by the Lyee theory.

The relation among the Lyee theory, Lyee methodology and Lyee program is shown in Fig. 2.

TDMs that is deduced from the conventional program by the analyzing program are shown in Fig. 3. This figure is called Process Route Diagram (PRD).

TDM expressed in a programming language is called Scenario Function (SF). PRD is expressed as SFs. SFs is a set taking SF as its elements, which are connected by Routing Vectors. The role of Routing Vectors is mentioned in the following chapter. Routing Vector is an element composing the algorithm for automatic conversion

B. Coordinates

2. Coordinates

A structure of a machine language , is broadly divided into two parts of operation and operand. There are two cases of the relation of these parts to user requirements:

(1) The relation between user requirements and the operation

(2) The relation between user requirements and the operand

The algorithm for automatic conversion utilizes these relations based on the Lyee theory. The relation (1) above is deterministically defined by a Predicate Structure. The role of the Predicate Structure will be discussed in the next chapter. The Predicate Structure is an element composing the algorithm for automatic conversion. The relation (2) is deterministically defined by coordinates. The role of the coordinates will also be discussed in the next chapter. The coordinate is an element composingcomposing the algorithm for automatic conversion.

Let us discuss the concept of the coordinates, which is used throughout this paper. To determine control logic based on user requirements is to define the relation (2) of the user requirements as a whole. However, we cannot define the whole but can merely agree at a certain level. Aforementioned meaning is defined on a basis of the whole, and we cannot define meaning since we are only able to define parts. Therefore, we use intention instead of meaning. Lyee methodology makes the operands correspond to words and makes the whole that we cannot define correspond to a network of words, i.e., a network of operands. However, this would be difficult work for humans as those who do this work should have empirical knowledge of semantic analysis. The outcome of the work should reflect the individual’s capability. The more ideas are put into the work, the more personalized is the outcome. Generally speaking, this causes lower productivity and lower maintainability.

In the Lyee methodology, a word network is created automatically instead of depending on individual capabilities, where SFs play an important role as mentioned later. The Predicate Structure is a concept realized in SFs and creates a word network automatically. The Predicate Structure is defined for each word. Each word takes coordinates. The coordinates are defined by SFs and Logical Behavior. That SFs and Logical Behavior can be derived from conventional programs by analyzing them makes the algorithm for automatic conversion be realized. The details of coordinates are mentioned later.

II. Principle of algorithm for automatic conversion

A. The role of TDM

The Three-Dimension-Like Space Model is shown in Fig. 4.

II. Principle of algorithm for automatic conversion (4 pages)

The role of TDM

Fig. 4

A word is a memorized entity. A set of words is considered to be user requirements. From the Lyee view point, a set of words is known. In conventional programs, words are defined as variable names in programs. Each word has a corresponding memory area. It is easy to take out these words from conventional programs using a program for this purpose. Logical Behavior is also taken out using a program for this purpose. The words are placed in the program, but in the Lyee theory, they are considered to exists in the natural space where memory is established. The cause to establish the words is in the consciousness space, a hypothetical space of the Lyee theory. We, of course, cannot memorize the state of the consciousness space. That is, the consciousness space is in the unknown state to us. We assume that a bridge between this unknown state and the memory could be built and the bridge is considered to be the intention. The bridge means correspondence of the consciousness space with the natural space. In order to realize the bridge, three sub-spaces are hypothesized. These five spaces and the relation among them are derived from the Lyee theory. The state memorized in the natural space and the state of the consciousness space, which is unknown to us, are both considered to have different characteristics while they are both existence. Existence in the sub-spaces bears the similar characteristics to that of the natural space whereas existence also inherits the characteristics unique to each sub-space. The existence mentioned above is assumed to have a structure. The structure consists of substance and attributes. The attributes are composed of a set of source of existence. The source of existence is called logical atom. The substance is composed of one logical atom. A pair of the substance and the attributes is referredconsidered to be existence. A word can beis considered as in memorized existence; hence, it has the structure mentioned above. A set of words are expressed in requirements, whereas words in the sub-spaces are also existence but not memorized.

The characteristics of the words in the sub-spaces are known and the words has the same structure as that of existence.

The

The existence in the consciousness space cannot be determined as a word, but its structure is considered to be the same. The characteristics of the existence in the consciousness space is known. These are all derived from the hypotheses of Lyee.

The word in the natural space and the characteristics of the existence in the three sub-spaces are known. The structure to point at the existence in the consciousness space using these known existence is TDM. Since the existence in the consciousness space cannot be pointed at and the characteristics of the structure as existence is known according to the hypothesis, we consider that when existence in the natural space and the sub-spaces are established simultaneously, the existence in the consciousness space is also established. Aforementioned synchronous state is expressed as the states being simultaneously established as mentioned here. In other words, this is similar to solve simultaneous equations. This means that the existence in the consciousness space is pointed at, which cannot be pointed. at. This also means that a correspondence of the known existence in the natural space to the unknown existence in the consciousness space is made. That is, the intention is established in our Lyee world.