Developing statistical software in Fortran 95.
COM server developmentcomponent object model standardconsole applicationsDLL FORTRAN routinesFORTRAN 95 programmingFORTRAN programming under Windowslogistic regression analysisnumerical accuracystatistical softwarestatistics environments interfacing with FORTRANstatistics programming developmenttextbook
Research exposition (monographs, survey articles) pertaining to statistics (62-02) Software, source code, etc. for problems pertaining to statistics (62-04) Research exposition (monographs, survey articles) pertaining to numerical analysis (65-02) Packaged methods for numerical algorithms (65Y15) Theory of programming languages (68N15)
The purpose of this book is to help statisticians, (theory and) application-oriented developers, and researchers that need statistical computer programs to write a good FORTRAN program for the first time. Since good programming style can be taught and learned, the authors try for their potential readers to get into the experience of developing FORTRAN applications as quick, painless, and rewarding as possible. The rapid growth of the computational statistics field is based on the large number of computationally intensive approaches to solving statistical problems and, accordingly, on the statisticians who need to begin programming for the development of statistical applications. The present book is addressed to the following large categories of target users: (a) Those accustomed to FORTRAN 77 and FORTRAN 90; (b) People wanting to move from FORTRAN console applications to the Windows environment; the book provides a step-by-step guide to separate the FORTRAN engine from the GUI (Graphical User Interface). (c) People wanting or needing to develop custom statistical routines that can be called from other applications, such as Excel, SAS, S-PLUS, SPSS, and MATLAB. The authors elaborated a GUI using the Visual BasicNET language, showing how computational routines written in FORTRAN can be packaged as DLLs (Dynamic-Link Libraries), and using the COM (Component Object Model) standard to interface object-oriented software with a COM server. The concise contents of the eight enclosed chapters of the book are as follows: The introductory Chapter 1 (Motivation) surveys the problems and the proposed solutions. Chapter 2 (Introduction to modern FORTRAN) and Chapter 3 (A pseudo object-oriented style) expose the basic programming aspects of the FORTRAN language, focusing on the new features of FORTRAN 90 and FORTRAN 95 standards. These modern features allow writing FORTRAN programs in a pseudo object-oriented style, expressed in a generic FORTRAN module developed for handling the run-time errors. Chapter 4 (Implementing computational routines) describes strategies for writing computational procedures while paying attention to issues of numerical accuracy, reliability, and efficiency. These principles of good programming design in FORTRAN are illustrated on creating a computational engine for logistic regression. Chapter 5 (Developing a console application) is a detailed guide for turning computational routines into reliable, platform-independent FORTRAN console program. The basic steps involved in a logistic regression analysis (loading the data, selecting model options, running the estimation procedure, examining the parameter estimates, and diagnostic measure) follow a pattern that is common to many statistical applications, serving as a template for many projects. Chapter 6 (Creating and using dynamic-link libraries) covers the creation and use of FORTRAN DLLs, as a basic mechanism by which a running program (written or not in FORTRAN) can communicate and access the functionality of a FORTRAN procedure. Chapter 7 (Creating COM servers) describes what COM servers are, demonstrates how to create them, and shows how the reader can turn a working FORTRAN console program into a COM server. Chapter 8 (Creating COM clients) discusses the process of building COM clients for statistical applications. There are presented clients such as a simple VBA (Visual Basic for Applications) script for a logistic regression analysis within Excel, as well as other clients in S-PLUS, SAS, SPSS, and MATLAB. An example of developing a stand-alone Windows application with the GUI written in Visual BasicNET is given. Finally, to notice that, for dissemination and educational purposes, all the resources mentioned within the book are available on-line at the Web site address: \texttt{http://methodology.psu.edu/fortranbook}.
- Dynamic link library for statistical analysis and its Excel interface
- Designing of the interface in the Visual C++ Environment for the program complex whose calculation part is compiled in Visual Fortran
- Fast analytical evaluation of intermolecular electrostatic interaction energies using the pseudoatom representation of the electron density. II: The Fourier transform method
This page was built for publication: Developing statistical software in Fortran 95.
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5463290)
