Transformations to Restructure and Re–engineer COBOL Programs

In this paper we describe a COBOL-program restructuring tool currently under development. The tool is constructed using program transformations executed by the TAMPR program transformation system. We discuss the COBOL knowledge embodied in the transformations and how they restructure an example COBOL program developed in the mid–1970s. While the tool is not yet a robust commercial product, early use for restructuring COBOL programs demonstrates the power and flexibility of this transformation–based approach.     

A structured COBOL

A version of COBOL that permits structured programming was designed. It was implemented using a pre-processor that outputs standard COBOL. The pre-processor will indiscriminately run programs that are either structured or non-structured, or with very limited restrictions, programs with a mixture of both. A structured IF was included as a modification of the COBOL IF. In line looping was created with a redefinition of the PERFORM, maintaining all of its capabilities and adding a DO UNTIL. The PERFORM was also modified to allow a CASE construct. All new structures are completely nestable.     

On program development effort and productivity

Many claims are made in the literature concerning large productivity increases as a result of the introduction of various new programming tools and techniques. Seldom do the researchers substantiate these claims, however. Further, many of the reported studies have been unsuccessful because much of the existing research is methodologically flawed and poorly grounded in theory. This paper investigates the methodological problems of previous studies and reports the results of a field study conducted to assess the feasibility of conducting controlled studies of the programming process. A field study of COBOL programs from three commercial organizations investigated the effects of programming style (i.e., structured programming) and programmer skill on (i) the effort required to develop programs and (ii) programming productivity; program size was used as a control variable. The results of the study support the underlying concept: that use of disciplined approaches and well-defined variables leads to more readily interpretable and more conclusive results.     

Measuring complexity in structured vs unstructured cobol programs

Measuring programmer productivity and estimating programming time and costs are among the most worrisome and persistent problems facing the programming manager. A key element in both problem areas is program complexity. It has been demonstrated in practice that a measure of program complexity is an indispensable aid in evaluating a programming effort. The purpose of this paper is to present a prototype for a composite measure of program complexity. The paper presents a basis for a technique upon which an objective quantitative evaluation for any program or programming effort could be made.This index of complexity would give the manager a tool for a quantitative assessment of programming efforts so that judgments about the relative merits of programs and programmers can be based on objective data and an objective measure. The measure is applied against a reference group of COBOL programs, several of which were written in a structured programming environment. The index of complexity and the data from which it is derived are used to evaluate the complexity of structured vs unstructured COBOL programming styles.     

Dynamic analysis of COBOL programs

This paper describes results obtained from a dynamic analysis of 22 typical COBOL programs. The analysis was performed by COBOL DYNAMIC ANALYSIS SYSTEM (COBDAP) which collected data and reported on the dynamic behaviour of COBOL programs.     

Cobol generation of source programs and reports

SURGE, a COBOL pre-processor for the machine generation of source programs, is described. The article presents the coding procedure by which the user communicates with SURGE, and briefly introduces the general logic of both the pre-processor and generated programs. The main capabilities of SURGE are described as file sorting, selective retrieval and tabular report preparation, including multiple levels of totals and a capability for certain other kinds of computation. Non-proprietary and relatively machine-independent, this package appears to reduce the need for hand-crafted COBOL programs and to be remarkably easy to learn and use. It is suggested that SURGE also may offer a suitable prototype for extending the concept of source-program generation to other goals such as file updating and the production of FORTRAN programs.     

Designing structured cobol programs

COBOL is now more than 20 years old and will probably survive to become much older. Since it has some features which are out of date it is desirable to adapt at least the program style to some standards of modern programming languages. The adaption is not only a matter of style but also of costs of program production and program maintenance. This paper presents constructional rules for programming in COBOL which by-pass some of the drawbacks and allow more readable and more maintainable program structures. Finally a postprocessor is presented, that allows verification of the chosen constructional rules and documentation of the resulting programs.     

COBOL到Java源代码翻译中的数据类型转换*

提出了一种功能等价的类型封装和嵌套方法,可以有效地将COBOL数据描述映射到Java类型系统.该方法已在所开发的COBOL2Java翻译系统中进行应用,通过了近400万行的真实银行商用系统的测试.实验结果表明,此方法正确有效,并且提高了生成代码的执行效率和可读性.     

A pilot project for migrating COBOL code to web services

This paper describes a pilot project conducted to test the feasibility of constructing web services from existing mainframe COBOL programs. The project involved the use of four tools. The first tool, COBAudit, was intended to identify candidates for web services. The second tool, COBStrip, served to extract only that portion of the code required to fulfill the service. The third tool, COBWrap, wrapped the code extracted from original code and converted it to an executable component. The fourth tool, COBLink, connected the wrapped component to the web by generating a WSDL interface from either the COBOL linkage section or the original map definition. The tools were applied to a legacy life insurance system with more than 20 million lines of COBOL code running under IMS on the IBM mainframe.     

Exception Handling in CLU

For programs to be reliable and fault tolerant, each program module must be defined to behave reasonably under a wide variety of circumstances. An exception handling mechanism supports the construction of such modules. This paper descnbes an exception handling mechanism developed as part of the CLU programming language. The CLU mechanism is based on a simple model of exception handling that leads to well-structured programs. It is engineered for ease of use and enhanced program readability. This paper discusses the various models of exception handUlng, the syntax and semantics of the CLU mechanism, and methods of implementing the mechanism and integrating it in debugging and production environments.     

On the performance of COBOL programs in large vs. mini computers

The comparative performance characteristics of COBOL programs in a small versus large computer systems are investigated. The vehicle consists of a set of synthetic benchmark COBOL programs, each measuring a particular aspect of COBOL computations, data manipulation, and input/output is made on both a large scale computer (IBM 370/158) and a minicomputer (Texas Instruments TI980). Results of a number of such experiments are presented and comparisons made between results obtained from the two systems.     

Computer education in times of recession: Should Pascal come first?

In recent years there has been a move towards making computer science courses more practical. In the current recession this seems even more important as employers can avoid candidates with an over-theoretical background.Pascal has become more popular in higher education as a first programming language to teach problem solving and structured programming. Structured programming is required in industry but there is evidence that the approach through Pascal is inadequate. Cognitive psychologists tell us that understanding the context of the problems to be solved is the key to problem solving. Pascal does not function within the commercial context. COBOL, on the other hand, does. By giving COBOL priority, students will gain an essential appreciation of the most common environment that they are likely to find themselves in at the end of their course. With the introduction of structured COBOL, they need not compromise on style either.     

Verification of Year 2000 conversion rules using the ACL2 theorem prover

The well-publicized Year 2000 problem provides interesting challenges for the remediation of noncompliant code. This paper describes some work done at EDS CIO Services, using the ACL2 theorem prover to formally verify correctness of remediation rules. The rules take into account the possibility of “flag” (non-date) values of date variables. Many of them have been implemented in an in-house tool, COGEN 2000^TM, that corrects for noncompliant date-related logic in COBOL programs.     

Elements of style among machine-generated COBOL flowcharts

After a brief review of the use of flowcharts both before and after the completion of a program, this article illustrates and compares the graphic output of four automatic flowcharting packages for COBOL programs. These packages are Autoflow, Dynachart, Quickdraw, and Axiom One. Primary consideration is given to the effectiveness of various notational conventions and schemes for geometric representation. The specimens presented may also suggest preferred practices for adoption in preparing hand-constructed documentation.     

Evaluating legacy system migration technologies through empirical studies

We present two controlled experiments conducted with master students and practitioners and a case study conducted with practitioners to evaluate the use of MELIS (Migration Environment for Legacy Information Systems) for the migration of legacy COBOL programs to the web. MELIS has been developed as an Eclipse plug-in within a technology transfer project conducted with a small software company [16]. The partner company has developed and marketed in the last 30 years several COBOL systems that need to be migrated to the web, due to the increasing requests of the customers. The goal of the technology transfer project was to define a systematic migration strategy and the supporting tools to migrate these COBOL systems to the web and make the partner company an owner of the developed technology. The goal of the controlled experiments and case study was to evaluate the effectiveness of introducing MELIS in the partner company and compare it with traditional software development environments. The results of the overall experimentation show that the use of MELIS increases the productivity and reduces the gap between novice and expert software engineers.     

An empirical analysis of COBOL programs

This paper describes results obtained from a static analysis of 340 COBOL programs collected from commercial and industrial installations. The analysis was performed by a syntax analyser designed specifically to analyse source program statements, gather detailed information and produce a report on the definition and use of data and language in the programs analysed.     

Formal methods reality check: industrial usage

Based on a systematic survey and analysis of the use of formal methods in the development of a dozen industrial applications, we summarize the methods being used, characterize the styles of industrial usage, and provide recommendations for evolutionary enhancements to the technology base of formal methods. The industrial applications ranged from reverse engineering to system certification; code scale ranges from 1 KLOC to 10 KLOC's. Applications included a software infrastructure for oscilloscopes; a shutdown system for a nuclear generating station; a train protection system; an airline collision avoidance system; an engine monitoring system for shipboard engines; attitude control of satellites; security properties of both a smartcard device and a network; arithmetic units; transaction processing; a real-time database for a medical instrument; and a restructuring program for COBOL     

An analytical study of strategy-oriented restructuring algorithms

Considerable experimental evidence has been accumulated showing that the performance of programs in virtual memory environments can be significantly improved by restructuring the programs, i.e., by modifying their block-to-page or block-to-segment mapping. This evidence also points out that the so-called strategy-oriented algorithms, which base their decisions on the knowledge of the memory management strategy under which the program will run, are more efficient than those algorithms which do not take this strategy into account.We present here some theoretical arguments to explain why strategy-oriented algorithms perform better than other program restructuring algorithms and determine the conditions under with these algorithms are optimum. In particular, we prove that the algorithms oriented towards the working set or sampled working set policy are optimum when applied to programs having no more than two blocks per page, and that, when this restriction is removed, they minimize both upper and lower bounds of the performance index they consider as the figure of merit to be reduced. We also prove that the restructuring algorithms aimed at reducing the page fault rate of programs to be run under such policies as LRU, Global LRU and PFF (the Page Fault Frequency policy) minimize an upper bound of the page fault rate, and we extend some of our results to some non-strategy-oriented algorithms. Throughout the paper, the only assumption about program behavior is that it can be accurately modeled as a stationary discrete-state stochastic process.     

Semi‐automatic grammar recovery

We propose an approach to the construction of grammars for existing languages. The main characteristic of the approach is that the grammars are not constructed from scratch but they are rather recovered by extracting them from language references, compilers and other artifacts. We provide a structured process to recover grammars including the adaptation of raw extracted grammars and the derivation of parsers. The process is applicable to possibly all existing languages for which business critical applications exist. We illustrate the approach with a non‐trivial case study. Using our process and some basic tools, we constructed in a few weeks a complete and correct VS COBOL II grammar specification for IBM mainframes. In addition, we constructed a parser for VS COBOL II, and were the first to publish a (Web‐enabled) grammar specification so that others can use this result to construct their own grammar‐based tools for VS COBOL II or derivatives. Copyright © 2001 John Wiley & Sons, Ltd.     

Composition of relational productions for plans and programs

A relational production system (rps) is a mathematical model for information processing in which computer programs and artificial intelligence plans have a common representation. Annth order composition theorem for relational productions is presented which specifies the net effect of a sequence of actions, providing a unified, operational, mathematical solution to the following problems: (1) formation of compound artificial intelligence operators; (2) closed-form representation of loop semantics for plans; (3) composition of program statements; (4) closed-form representation of loop semantics for programs. This leads to a composition-based verification method, in which structured data are treated by the same methods as unstructured data. This approach seems especially promising in the verification of programs with structured data, which the orthodox inductive assertion technique accommodates only after major transfusions of additional theory.