软件工程基础：抽象与建模原理探索

5星 · 超过95%的资源需积分: 15 50 浏览量更新于2024-07-30 1 收藏 53.82MB PDF 举报

"《软件工程卷1：抽象与建模》是Dines Björner等编著的一本专业书籍，属于Texts in Theoretical Computer Science系列，由Springer出版。该书深入探讨了软件工程中抽象与建模的重要概念和技术。书中首先介绍了离散数学的基础，包括数、集合、函数等概念，接着讲解了面向属性和面向模型的规约原理，特别通过RAISE规约语言RSL来阐述这些概念。此外，书中还涵盖了应用式、命令式和并发式规约程序设计的基本原则，并提供了全面的软件工程术语表和丰富的索引及参考文献，旨在为读者提供全面的软件工程理论基础。" 本书的重点在于软件开发中的抽象和建模技术，这是软件工程中的核心部分。抽象帮助开发者将复杂的问题简化为更易于处理的部分，而建模则是将现实世界的问题转化为计算机可理解的形式。书中详细阐述了离散数学作为这些概念基础的重要性，离散数学是计算机科学的基石，它的概念如集合论、函数和逻辑在软件设计中有着广泛的应用。在规约方面，作者介绍了不同类型的规约语言，如B、VDM-SL和Z，但主要通过RAISE规约语言RSL来具体讲解。RAISE是一种形式化方法，用于规格说明和系统验证，其面向模型的特性使得它在软件工程中具有较高的实用价值。同时，书中还讨论了应用式（函数式）、命令式和并发式（并行式）编程范式的规约设计，这些都是现代软件开发中常见的编程风格。书中的内容不仅限于理论，还包含了38个图表，有助于读者直观理解复杂的概念。全面的软件工程术语表和详尽的索引则为读者提供了查阅和深入研究的便利。此外，参考文献列表为那些希望进一步探索特定主题的读者提供了宝贵的资源。《软件工程卷1：抽象与建模》是一本适合软件工程师、计算机科学学生以及对软件开发理论感兴趣的人士阅读的专业著作，它提供了一个全面而深入的视角，帮助读者理解和掌握软件开发中的关键概念和技术。

XVI Preface — to Vols. 1-3

safety-critical real-time software, database systems, etc., to become a proper

specialty software engineer.

The Author's Aspirations

So these then were and are my aspirations: To provide you with a differ-

ent kind of textbook; to bring more than 30 years of exciting programming

methodological studies and controlled experimental practice into the larger

arena of software engineering; to show you what a beautiful world software

development can be when following the didactic cornerstones of linguistics,

philosophy, semiotics and mathematics; and to unload more than 25 years of

evolving lecture notes into a set of three coherent, consistent and relatively

complete volumes.

I have written these volumes because I wanted to understand how to

develop large-scale software systems. When I started, some 25 years

ago,

writing lecture notes on this subject, I knew less than I do now.

Meanwhile I have had the great pleasure of having many clever and

eager students follow the practice. I have initiated the large-scale com-

mercial developments of compilers for such unwieldy programming

languages as CHILL [254,255] and Ada [128,129,155], and I have thus

honed and corrected my thinking. Writing about software engineering

while testing out the ideas has been a sobering experience. There are

still many corners of software engineering that I have to write about,

think and experience. Meanwhile, this is what you get!

These volumes thus represent my chef d'ceuvre.

Role of These Volumes in an SE Education Programme

Who are the target readers of these volumes? That question is indirectly

answered in the following.

What roles do we see these volumes serve in the larger context of an

academic software engineering education, one that leads to a Master's degree

in the subject? Figure 1 shall assist us in answering that question.

The labelled boxes of Fig. 1 designate topics that enter into the software en-

gineer's daily practice, and which are therefore useful topics of learning. In Fig. 1

two-way arrows between boxes indicate that the designated topics can be studied

simultaneously. Directed (one-way) arrows between boxes designate a suitable, pro-

posed precedence relation between the learning of these topics. A "fan in" (multiple

source) arrow shows that a topic may need (i.e., have as prerequisites) the knowledge

of one or more (predecessor) topics. A "fan out" (possibly multiple target) arrow

shows that the arrow source topic is a "must" for one or more successor topics.

Preface — to Vols. 1-3 XVII

Mathematical

Logic (1)

Imperative

Programming

(Java)

SOFTWARE ENGINEERING

Volume 1

Basic Abstraction Principles

Recursive

Function

Theory

Functional

Programming

(Standard ML)

Mathematical

Logic (II)

Logic

Programming

(Prolog)

Algebra

Parallel

Programming

(occam)

SOFTWARE ENGINEERING

Volume 2

Advanced Specification Techniques

SOFTWARE ENGINEERING

Volume 3

From Domains via Requirements to Software

SOFTWARE ENGINEERING

suitable Text on

Software Engineering Management

Fig. 1. Courses based on these volumes: a first setting

We emphasise that we here place these volumes in the context of an academic

Software Engineering MSc education programme — not to be confused with an

academic Computer Science MSc education. The former aims at the produc-

tion of industry programmers: developers of commercial software. The latter

aims at theoreticians, useful in an academic institution of study. Another ex-

planation, wrt. another diagram, would thus have to be given for an equally

likely setting in the context of an academic programme for an MSc degree

in (theoretical) computer science, and yet another one for an undergraduate

course of an academic software engineering BSc education programme.

• Prerequisite or "concurrent" courses: We assume that the reader of

these volumes is — or while following a course based on Vol. 1 of these

volumes becomes — familiar with the general topics of imperative, func-

tional, logic, parallel and machine programming. Teaching in these topics

must cover both skill-learning and training wrt. specific languages such as,

for example, SML (Standard ML) [261,389] for functional programming,

Prolog [295,351] for logic programming, Modula-3 [262,401], Oberon [527]

and Java [10,20,243,348,470,511] for modular (i.e., object-oriented) pro-

gramming, and occam [364] and a machine language for some well-chosen,

"current-technology" hardware (e.g., Intel-like) chip. Their teaching must

also cover — to a basic extent — the knowledge acquisition wrt. the theo-

retical background for these programming styles and languages: recursive

function theory [136,444], logic for logic programming [295,351], Hoare

Logic for imperative programming [15,16] and process algebras for concur-

rency (CSP [288,289,448,456] and Petri nets [313,421,435-437]). The ma-

chine programming topic [379,501,511] is the only real hardware-oriented,

but not hardware-design-oriented [279,418], course. Codesign

[482],

that

XVIII Preface — to Vols. 1-3

is,

design of combined hardware/software systems (typical, for example,

for embedded systems, see below) is not covered. But one could "add

other boxes"! Included in the above kinds of course, or additional to these,

we expect the reader to have some working knowledge of algorithms and

data structures, i.e., to be familiar with the classical as well as modern

such algorithms and data structures and measures of concrete complex-

ity [7,357,371,495,524].

Auxiliaries: The reader is assumed to be — or to become, in conjunc-

tion with the software engineering study of which these volumes are part

— comfortable with mathematics — to a Bachelor's degree level in the

subjects listed. We suggest

[534],

a delightful "smallish" introduction, and

the substantial introduction to discrete mathematics

[213].

We find [213]

to be an excellent textbook for an entirely separate, and major, course on

that topic. One that every software engineer is assumed to take.

Software

Tools Laboratory

Unix/Linux

Windows

GUI,

Graphics, Scripts

Web/Internet

Lexical Scanners

Parsers

Etcetera

Algorithms

Data Structures

Machine

Programming

SOFTWARE ENGINEERING

Volume 1

Basic Abstraction Principles

SOFTWARE ENGINEERING

Volume 2

Advanced Specification Techniques

SOFTWARE ENGINEERING

Volume 3

From Domains via Requirements to Software

[

SOFTWARE ENGINEERING

A suitable Text on

Software Engineering Management

Graph Theory

and

Combinatorics

Calculus:

Differential

Equations

Statistics

and

Probability Theory

Operations

Research

Fig. 2. Courses based on these volumes: a second setting

Similarly, but more thought of as part of term projects and other forms

of laboratory (including self-study) work, we expect the reader to be rea-

sonably comfortable with practical, existing platform technologies (the

Software Tools Laboratory box).

Main course: These volumes are then to serve in a main set of three

courses on software engineering — and before the breadth and depth of the

follow-on courses are attempted. We additionally would advise acquisition

of the two books [236,238], the first as supplementary, the second to fill

out especially the verification (i.e., the design calculi) parts which are not

developed in these volumes.

Preface — to Vols. 1-3 XIX

• Follow-on courses: Classical software engineering has focused rather

much on the "navel-gazing", i.e., introspective parts of computing systems:

compilers, database systems, distributed systems, operating systems, real-

time (fault-tolerant and) embedded systems, etc. Ideally such topics should

now be covered on the basis of, and from the point of view of, formal spec-

ification and design calculi. The embedded systems topic (given 1-3 units

of work load) could go as far as including hardware/software codesign [482]

and otherwise rely strongly on other systems engineering issues.

Linguistics of

Programming

Languages

Semantics of

Programming

Languages

Compiler &

Interpreter

Development

SOFTWARE ENGINEERING

Volume 1

Basic Abstraction Principles

SOFTWARE ENGINEERING

Volume 2

Advanced Specification Techniques

SOFTWARE ENGINEERING

Volume 3

From Domains via Requirements to Software

SOFTWARE ENGINEERING

A suitable Text on

Software Engineering Management

Systems

Programming

Practicum

Application

Systems

Practicum

Operating

Systems

Development

Embedded,

Safety

and Real-time

Critical Systems

Database

Theory and

Systems

Distributed

Systems

Database

Management

Systems

Cryptography

and

Secure Systems

Fig. 3. Courses based on these volumes: a third setting

Additionally we plead that each software engineering student take two

"practica": A large, project and colloquium/seminar-oriented "systems

programming" and a similar "applications systems" course which exper-

imentally and exploratively researches and develops a non-trivial hard-

ware/software control system, respectively a commercial, industrial, or

other application such as amply hinted at in these volumes!

• A final software engineering course is hinted at: "Software Engineering

Management". We have quite some material for lectures on this topic. For

the time being we refer to the excellent book by Hans van Vliet

[512].

The linguistics of formal languages, including theories of formal semantics, is

crucial knowledge to be possessed by the professional engineer. Two courses

relate to this: the leftmost boxes in rows three and four of Fig. 3. The lin-

guistics course could be based, for example, on David A. Schmidt's or John

Reynolds' works. References are

[455],

respectively

[442].

The semantics course

could be based, for example, on any of [183,252,443,454,497,521].

XX Preface — to Vols. 1-3

Why So Much Material?

These volumes are more-or-less self-contained. We expect these volumes to be

used in university and college courses, and to be studied by readers on their

own. Some universities and some colleges cover material in courses that lie

early in the course curriculum that we also bring here. So it could be assumed,

and left out? No, not quite, since other universities and colleges do not cover

such prerequisite material. Hence these volumes, again and again, must make

these excursions. Since these volumes significantly rely on mathematics —

not anything advanced, not something for which any deep theorems need be

known or used — we need to recapitulate some of this material in Chaps. 3-9.

There we also explain and illustrate the A-calculus.

Since actual life phenomena have to be perceived, whether manifest or

not, that is, have to be conceptualised, we make deep excursions, in Vol. 3's

Chaps. 3,5,6 and 7, into what constitutes a methodology, what are definitions,

what are phenomena and concepts, and what is a description.

Since language is such an important basis for all we do in software en-

gineering, and since we cannot rely on the necessary topics having already

been learned, i.e., being known, we need also to make deep excursions into

the pragmatics, the semantics and the syntax, in toto, the semiotics of lan-

guages, whether formal or informal. Since automata and finite state machines

likewise form an indispensable component of our science and engineering we

need also cover that topic in Vol. 2's Chap. 11. In covering all these adjunct

ideas we supply their treatment with a twist: We present them from unconven-

tional angles. We expect, thereby, that the reader achieves a different view on

these matters, one that is more relevant to engineering than perhaps science,

more relevant to practice than to theory. In any actual course the lecturer can

therefore, based on local course curricula, leave out some of the "excursion"

material.

How to Use These Volumes in a Course

Together with these volumes it is planned to make available over the Internet:

• http://www.imm.dtu.dk/~db/The-SE-Books

a comprehensive set of electronic documents:

• a large variety of suggested course structures (with references to volume

chapters and slides)

• group project descriptions — some with solutions

• large-scale development examples

• URLs to formal methods pages

• URLs to formal methods tools

剩余734页未读，继续阅读

lhclylzl

粉丝: 1
资源: 4

软件工程基础：抽象与建模原理探索

软件工程卷1：抽象与建模(1/3)

医学统计学山东大学课程

夜神模拟器

unix网络编程 卷1:套接字联网api

软件工程期末复习试卷

ia-32 intel 架构软件开发人员手册卷 3:系统编程指南 type:pdf

ABAQUS软件中如何设置并执行一个线性静态分析，并详细解读应力分布结果？

创建一个数据卷容器1和一个数据卷1， 将数据卷挂载到数据卷容器 在数据卷中创建文件， 然后备份该数据卷。 然后将该备份文件恢复到创建的容器2中的数据卷2中。

如何运用UML图表详细建模网络考试系统？请结合《UML建模：网络考试系统分析与设计》进行详细说明。

java核心技术 卷ii:高级特性(原书第10)

最新资源

unix网络编程卷1:套接字联网api

创建一个数据卷容器1和一个数据卷1，将数据卷挂载到数据卷容器在数据卷中创建文件，然后备份该数据卷。然后将该备份文件恢复到创建的容器2中的数据卷2中。

java核心技术卷ii:高级特性(原书第10)