Software engineering
zh-cn:软件工程 Software engineering is the technologies and practices used to create and maintain computer software, while emphasizing productivity and quality. In the year 2000, these technologies and practices encompass languages, databases, tools, platformss, libraries, standards, patterns, and processes.Software engineering applications include email, embedded software, graphical user interfaces, office suites, operating systems, optimizing compilers, relational databases, robotics controllers, video games, and the world wide web. Other important applications include accounting, airline reservations, avionics, banking, and telephony. These applications embody social and economic value, in that they make people more productive, improve their quality of life, and enable them to do things that would otherwise be impossible.
Software engineers are the community of practitioners who create programss. In the year 2000, there were about 640,000 software engineers in the U.S and another (1,200,000?) in Europe, Asia, and elsewhere. Steve McConnell says 40% have computer science degrees. Many others have science, engineering, business or other, non-technical degrees. They use diverse technologies and practices and work on diverse applications.
Software engineering today
Software engineering is alive and well today.
Importance of software engineering
In the USA, software drove 1/4 of all increase in GDP during the 1990s (about $90 billion per year), and 1/6 of all productivity growth (efficiency within GDP) during the late 1990s (about $33 billion per year).
Software engineering has changed the world's culture, particularly in countries whose people use computers. Email, the world-wide web, and instant messaging enable people to interact in new ways. Software lowers the cost of many important social goods and services, such as health-care and fire departments.
Examples of successful large projects, where software engineering methods have been applied, include the space shuttle software. If it is cheaper to run a business with application software than to forego the cost of doing business with computing systems, a business may choose to sink funds into computer hardware, software and personnel.
Current directions for software engineering
Agile processes are an important emerging practice. Agile processes are methods to manage software development so projects evolve with changing expectations. The older document-driven processes (like CMM and ISO 9000) may be fading in importance. Some persons believe that companies have exported many of the jobs that can be controlled by these processes. Related concepts are Lean software development and Extreme programming.
Aspect programming is also an important emerging programming technology. Aspects help programmers deal with ilities by providing tools to add or remove boilerplate code from many areas in a software project. Aspects describe how all objects or functions should behave in a particular circumstance. For example, aspects can add debugging, logging, or locking control into all objects of a particular type. Related concepts are Generative programming and Templates.
The future of software engineering was an important conference at the ICSE 2000, [1]. The FOSE project summarized the state of the art in 2000 and listed many problems to be solved over the next decade. The Feyerabend project is attempting to discover the future, [1].
Today's Practice
In general, few programmers, much less software engineers, are free agents, but most work as employees or contractors (for others).
A sub-community of computer-related personnel indeed requires the expertise of computer specialists, and the computer specialists share common language, terminology, certification, and so forth. So, the specialists can be considered part of a larger software, hardware and application community.
Evolution
This section discusses the evolution of software engineering.
1945 to 1965: The origins
The term software engineering first was used around 1960 as researchers, management, and practitioners tried to improve software development practice. Programmers were well aware of electrical and computer engineering and debated what software engineering might mean.
The NATO Science Committee sponsored two conferences on software engineering in 1968 (Garmisch, Germany) and 1969, which gave the field its initial boost. Many consider these conferences to mark the start of the profession.
1965 to 1985: The software crisis
Software engineering was spurred by the so-called software crisis of the 1960s, 1970s, and 1980s, when many software projects had bad endings. The software crisis was the name that lumped together the many problems with software. Many software projects ran over budget and schedule. Some projects caused property damage. A few projects caused loss of life. The software crisis was originally defined in terms of productivity, but evolved to emphasize quality.
Cost and Budget Overruns: The OS 360 operating system was a classic example. It is still used on the IBM 360 Series and its descendants. This decade-long project eventually produced a working system amongst the most complex software systems ever designed. OS 360 was one of the first very large (1000 programmer) software projects. Fred Brooks admits in Mythical Man Month that he made a multi-million dollar mistake when he managed the project.
According to a study by the Standish Group, in 2000, only 28 percent of software projects could be classed as complete successes (meaning they were executed on time and on budget), while 23 percent failed outright. Many projects were abandoned.
Property Damage: Software defects can cause property damage. Poor software security allows hackers to steal identities, costing time, money and reputations. The explosion of a European Ariane rocket and other disasters spurred further developments in the field.
Life and Death: Defects in software have killed. Several embedded systems used in radiotherapy machines failed so catastrophically that they administered lethal doses of radiation to patients.
A large list of conmporary software problems and disasters is kept at Peter G. Neumann's Computer Risks column.
1985 to 1995: No silver bullet
For decades, solving the software crisis was paramount to researchers. Seemingly, every new technology and practice from the 1970s to the 1990s was trumpeted as a silver bullet to solve the software crisis.
Technologies: Every new technology and practice for decades was touted as the solution to the software crisis: structured programming, object-oriented programming, process, CMM, UML, Ada, methodologies, and so on.
Practices: Some pundits argued that the software crisis was due to the lack of discipline of programmers. Some believed that if formal engineering methodologies could be applied to software development, the production of software would become as predictable an industry as other branches of engineering. This led to work on a code of ethics, and professionalism. Many practitioners resist process strongly.
In 1987, Brooks published the famous paper "No Silver Bullet", arguing that no individual technology or practice can make a 10-fold improvement in productivity in 10 years. See his article
Debate about silver bullets raged over the following decade. Some advocates for particular technologies continued to argue that their favorite technology or practice would be a silver bullet. Sceptics disagreed. But eventually, most everyone accepted that no silver bullets would ever be found. Yet, the term silver bullet pops up now and again, even today.
All known technologies and practices make incremental improvements in productivity and quality.
1995 to present: Emergence as a profession
In the mid-1990s to mid-2000s, software engineering emerged as a bona fide profession, to stand beside computer science and traditional engineering.
Prior to the mid-1990s, most software practitioners called themselves programmers or developers, regardless of their actual jobs.
The term programmer had often been used as a pejorative term to refer to those who lacked the skills, education, or ethics to write quality software. Practitioners began to describe themselves as software engineers to escape the stigma attached to programmer. In many companies, the titles "programmer" and "software developer" were changed to "software engineer", for many categories of programmers.
This caused confusion, because some denied any difference while others tried to create a difference. Traditional engineers questioned whether software engineers could legally use the term.
Notes
The field has evolved steadily.
Practice: There has been a continuity of practice from the 1940s to the 2000s.
Relative Cost: The relative cost of software versus hardware has changed substantially. When mainframes were expensive, software projects could be expensive. Because powerful PCs are cheap, software costs must become cheaper, in comparison.
Women's issues
In the 1940s, 1950s, and 1960s, software was a women's ghetto. Men preferred the higher prestige of hardware engineering roles. So, women filled the programming roles. Women like Grace Hopper were common.
Many unsung women wrote code prior to 1968. Today, relatively few women work in software engineering. Women have largely moved into analysis and testing roles. Saying that this is sexual discrimination is too simple, because it related directly to individual identity. In this sense, software engineering is the masculinization of programming.
The role women play continues to evolve. Today, women in software engineering frequently do analysis, testing, education, documentation, and management, rather than hard-core development.
Ada Lovelace was deeply in debt, and was attracted to Babbage's scheme so that she could gamble on horses more effectively.
Comparing Related Fields
The relationship between software engineering and the related fields of programming, computer science, and traditional engineering has been debated for many decades. Many of the similarities and differences are discussed here.
Whether software development is more like art, science or engineering has been fiercely debated for many decades. Software development shares attributes of all of these fields, and many software projects have elements of all, but important distinctions exist.
Comparing programming
Programmers emphasize the task of writing code. Software engineers work on all sizes of applications: small and large.
Programming emphasizes coding to produce working software applications, independent of budget and schedule.
Software engineering tries to encompass software projects more completely, including budget and schedule. Software engineering recognizes that software development fits in a large business context with relationships to marketing, sales, production, installation, training, support, and operations. Software engineering emphasizes methods to construct large applications that individual programmers cannot write alone. Software engineering tries to come up with methods to write programs in a consistent way.
Comparing computer science
Many believe that (metaphorically) software engineering is to computer science and information science as traditional engineering is to physics and chemistry.
While 40% of software engineers earn computer science degrees, they do not practice computer science every day, which is different than practicing software engineering every day.
| Issue | Software Engineering | Computer Science |
|---|---|---|
| Ideal | Constructing software system for real-world use | Correct eternal truths about computability and algorithms with good runtime behavior |
| Goals | Working programs (like office suites and compilers) | Algorithms (like Shell sort) and abstract problems (like Travelling Salesman) |
| Budgets and Schedules | Projects (like the next upgrade) have fixed budgets and schedules. | Projects (like solving NP) are independent of budget and schedule. |
| Emphasis | Software engineering emphasizes applying skills and working programs that deliver value to users. | The theory of software emphasizes eternal truths, like the running time analysis, space analysis, and correctness of algorithms. |
| Change | Programs will evolve as user needs and expectations evolve, and as technologies and practices evolve. | When computer science problems are solved, the solution will never change. |
| Additional Skills | Domain knowledge | Mathematics |
| Notable Educators and Researchers | Barry Boehm, David Parnas, and Frederick P. Brooks | Edsger Dijkstra, Donald Knuth, Robert Tarjan, and Alan Turing |
| Notable Practitioners | John Backus, Dan Bricklin, Tim Berners-Lee, Linus Torvalds, Richard Stallman | Not applicable |
| Number of Practitioners in U.S. | 640,000 | 25,000 |
| Number of Practitioners in World | unknown | unknown |
Comparing traditional engineering
Some practitioners believe that they apply concepts of traditional engineering to software development. They believe engineering provides a structured, logical approach, and therefore, a stable final product. Other practitioners are inspired by traditional engineering, but believe that software problems need particular solutions. They believe that traditional engineering concepts may not apply, because software is fundamentally different than bridges and roads. For example, traditional engineers do not use compilers or linkers to build roads.
Software engineers aspire to build low-cost, reliable, safe software, which is much like what traditional engineers do.
Software engineers borrow many metaphors and techniques from traditional engineering disciplines: requirements analysis, quality control, and project management techniques.
Traditional engineers now use software tools to design and analyze systems, such as bridges and buildings. These new kinds of design and analysis resemble programming in many respects, because the work exists as electronic documents and goes through analysis, design, implementation, and testing phases, just like software.
Few engineering fields have not included some link to computing. For example, if it is cheaper to execute a computer simulation than to implement an engineering model, a traditional engineer will be pressured to create a computer simulation.
| Issue | Software Engineering | Traditional Engineering |
|---|---|---|
| Foundations | Software engineering is based on computer science, information science, and discrete math. | Traditional engineering is based on physics, chemistry, and calculus. |
| Cost | Compilers and computers are now cheap, so software engineering and consulting often cost more than 50% of a project. Even minor software engineering cost-overruns can cause a project to fail. | Construction and manufacturing costs are high, so traditional engineering may only cost 15% of a project. Even major engineering cost overruns may not affect a project's viability. |
| Management Status | Few software engineers manage anyone, so they are not viewed as managers, except by themselves. | Many traditional engineers manage construction, manufacturing, or maintenance crews, so they are all treated as managers. |
| Innovation | Software engineers apply new and untested elements in many software projects. | Though some projects have innovations, traditional engineers apply known and tested principles, and limit the untested innovations that goes into each product. |
| Replication | Replication is trivial, and most development effort goes into building new (unproven) or changing old designs and features. | Most development effort goes into replicating proven designs. |
| Number of Practitioners in U.S. in 2000 | 640,000 | 1,100,000 total engineers 65,000 computer engineers |
The term engineering causes a lot of confusion. Some believe that it means that practitioners must be engineers. Others believe that software engineering is a metaphor. Engineers may work with a leading agency, be it a profit-seeking corporation (a business), a government (civilian or military), or non-profit agency such as a school or .org, such as Wikipedia.
In the U.S., there are 10 times as many software engineers as computer engineers, and the software engineering community is about 60% as large as the traditional engineering community.
Software engineering is much younger than many traditional engineering disciplines such as civil engineering.
Some claim that software engineering is already as predicatable and reliable as many fields of engineering, such as space or biological engineering. Although large, reliable software systems can be and have been constructed, software projects that fail during construction or in service are still too common. However, failures in large traditional engineering systems, such as those preceding the disasters in Three Mile Island, Chernobyl, Bhopal Disaster, Space Shuttles Challenger and Columbia are also too common. Others argue that unlike in traditional engineering where practicioners analyze failures, find precise causes, and set up guidelines to avoid them in the future software engineers routinely fail to pinpoint causes of failure or delay precisely enough to avoid repeats in the future.
Differences of Opinion
As software becomes more pervasive, we all recognize the need for better software. Yet, proponents advocate conflicting solutions. Different methodologies advocate different solutions. Proponents of different methodologies often get into heated debates over their merits.
Everyone agrees that we want better software. We disagree on priorities and approach, on what an individual should do first in a specific circumstance.
With an industry of 640,000 software engineers and 530,000 programmers in the USA, there should be room for many approaches. Subfields like consumer applications are very sensitive to time and cost. Subfields like military and medical applications are very sensitive to quality. All subfields mix these needs to varying degrees. There should be room for people to try different approaches. However a consensus has yet to emerge.
The usage of these technologies, practices, and applications can serve to define who precisely belongs to the software engineering community. However, different people argue for different criteria.
Fighting over words
Traditional engineers (especially civil engineers) claim that they have a special claim to the term engineering, and for any field to use the term requires their approval.
The fields of data engineering, knowledge engineering, user interface engineering, and so on have similar concerns over engineering issues.
Many people prefer the term software developer, though some still prefer the term "programmer". Many argue that they all essentially do the same thing with computers.
Fighting over issues
Everyone seems to emphasize a different combination of the following issues.
Management practices: Some argued that software engineering was primarily about the management practices necessary to make budgets and schedules. The Software engineering institute took this approach to create the CMM.
Formal methods: Some applied rigorous mathematical analysis to computer programming. The perception at the time was that conventional engineering was carried out with a great degree of mathematical rigor, while computer programming was commonly seen as an iterative trial and error process. It's likely that neither of these perceptions were especially accurate, but the term software engineering was adopted to indicate an intent to make programming more rigorous.
Methodologies: Some have argued that software engineering must follow a step-by-step methodology, much like an assembly line. This inspired many methodologies, such as RUP.
Tools: Many took the approach that software engineering must mean tools, especially CASE tools.
Traditional engineering: Many have tried to define programming as a subset of traditional engineering. This means that software engineers should study physics and chemistry.
Professionalism: Some have interpreted software engineering in terms of codes of ethics and state licenses, like traditional engineers have.
Criticisms of software engineering
Critics charge that some assumptions made during the process of implementing software engineering are inherently flawed. However, these criticisms may seem to apply to every human activity (Eg : fine arts or even quantum cryptography). Hence, a few believe that these criticisms do not hold much water. The following paragraphs detail many of the criticisms and some responses to them.
- Managing Expectations
- Criticism: One key to successful "software engineering" projects is managing the customer's expectations to something that can be built and delivered. So, the field resembles marketing, sociology, or voodoo, more than traditional engineering with its responsibilities to society at large and the perils of legal liability with failure to protect the public interest.
- Response: Every profession manages expectations, including all forms of engineering. This may happen at different levels. Software engineering focuses on immediate requirements, whereas other engineering fields tend to solve problems that long-ranged (bridges and dams). Moreover, "responsibility to the society" means meeting the expectations of the general public, which can be considered a customer.
- Poor Requirements
- Criticism: The requirements for most projects are incomplete or inconsistent. Some customers have little previous experience in writing requirements for a software project. Other customers really do not know what they want, and say "I'll know it when I see it" (IKIWISI). Even experienced customers who know exactly what they want may be unable to articulate their requirements. Users frequently expect far more than they write in specifications. And, complete requirements may describe programs that have no computable or practical solutions.
- Response One response to this objection is that one must avoid all unclear projects (but there might be no others left). Another response is agile development and rapid prototyping to clarify project goals as soon as possible. Some branches of software engineering like Embedded systems are unique in that they are designed by other engineers, who can sometimes define interfaces completely.
- Rising Complexity
- Criticism: Critics argue that the probability of failure increases with the size, scope and complexity of the project. Technologies and practices have improved through the 1970s, 1980s, and 1990s, but the complexity in requirements and user expectations have also increased fast enough to overwhelm the improvement in practices. So, the gap between what is expected and what is delivered has not improved.
- Response: This is not the fault of practitioners. This is actually a measure of the success of practitioners, because demand normally follows the supply. So customers demanding more is an obvious sign of their belief that what they demand will be supplied (even if they peg the demand slightly above the measure of supply in order to enforce agility.)
- Ongoing Change
- Criticism: Practitioners are eager to develop new technologies and practices and try new tools in every project. Some view this ongoing change as proof that older technologies and practices were failures.
- Response: Many view this ongoing change as proof that software engineering successfully learns and grows.
- Ongoing Failure
- Criticism: Critics charge that incomplete or poorly designed systems are still too common. The early lessons of the field were not sufficient to prevent subsequent disasters.
- Response: No field that strives to do bigger and better projects has ever avoided all unintentional failures and compromises. Moreover, traditional engineering fields frequently resort to a balance of factors to achieve optimization, and hence this is not an isolated problem.
- Nothing New
- Criticism: Some argue that software engineering has not created anything on its own, it merely uses what computer scientists already know.
- Response: Practitioners developed many tools and practices (compilers, make, cvs, XP), on their own, out of simple need. These tools make use of the tenets of computer science, just like the other engineering fields that make use of pure sciences like math and physics.
- Anyone Can Do It
- Criticism: Many bright people from other fields (engineers, scientists, business people) write spreadsheet templates or calculator programs, and eventually switch to writing large applications, and believe that they do software engineering. So, software engineering is not a special skill.
- Response: Software engineering is a skill that is refined through long practice. Software engineers are the ones who already have the education and experience, and they keep up with evolving technologies and practices. This is true of every skill.
- We Do Not Know What It Is
- Criticism: It doesn't yield to the standard ways of categorization, under accepted definitions of engineering. This is usually claimed by someone who wants to impose his/her own definition.
- Response: We know a lot about what software engineering is. Software engineering is grounded in the technologies and practices, applications, and the community of software engineering practitioners.
- Software as Executable Knowledge
- Criticism: In Zepplins and Jet Planes [1], Philip Armour argues that software is executable knowledge, meaning that the development of this knowledge is a major part of software development. Knowledge is discovered in an creative process where trial and error, learning, and the ability to challenge one's assumptions are important. In this view, the potential benefit of software engineering is limited to making it easier for developers to trial ideas, to discover errors earlier, and to provide them with information about the state of a software system.
- Criticism: In Zepplins and Jet Planes [1], Philip Armour argues that software is executable knowledge, meaning that the development of this knowledge is a major part of software development. Knowledge is discovered in an creative process where trial and error, learning, and the ability to challenge one's assumptions are important. In this view, the potential benefit of software engineering is limited to making it easier for developers to trial ideas, to discover errors earlier, and to provide them with information about the state of a software system.
Methodologies
Software engineering methods span many disciplines, including project management, analysis, specification, design, coding, testing, and quality assurance. All of the methods guiding this field are collations of all of these disciplines. Software design methods can be informally classified into thick and thin.
Thick methods include a large amount of formal process paperwork and documentation. Well-known thick methods include Cleanroom, ISO 9000, CMM and Rational Unified Process (RUP).
Thin methods eschew formal process paperwork and documentation. Well-known thin methodologies include Extreme Programming (XP) and Agile Processes.
Recently, some (like Karl Weigers) have argued for no more methodologies. Methodologies tend to list the contemporary technologies and practices and insist that everyone use them. This advice is obvious for those who work on new systems and have the opportunity to use contemporary technologies and practices. This advice is useless for those who maintain legacy systems and must use legacy tools and must use older technologies and practice, due to circumstance. So, methodologies are not specifically useful. And methodologies must be updated as technologies and practices evolve.
Processes and meta-processes
There is a growing body of software development organisations that are implementing process methodologies. Many of them are in the defence industry, which in the U.S. requires a 'Rating' baised on 'Process models' to obtain contracts.
The Capability Maturity Model (CMM) grades organizations on how well they create software accrording to how they define and execute their processes. ISO 9000 describes standards for formally organizing processes with documentation.
ISO 15504 or SPICE, The software process life cycle is also gaining wide usage. This standard is aimed at setting out a clear model for process comparison. SPICE is used much like CMM and CMMI. It models processes to manage, control, guide and monitor software development. This model is then used to measure what a development organization or project team actually does during software development. This information is analyzed to identify weaknesses and drive improvement. It also identifies strengths that can be continued or integrated into common practice for that organization or team.
Six Sigma is a methodology that uses data and statistical analysis to measure and improve a company's operational performance. It works by identifying and eliminating "defects" in manufacturing and service-related processes. The maximum permissible defects are 3.4 per million opportunities. However Six Sigma is manufacturing-oriented, not software development-oriented and needs further research to even apply to software development.
Extreme programming, Agile processes, and Lean software development are full blown processes that take an incremental or evolutionary approach to software development.
Professional Issues
For many years, software engineering has been trying to call itself a profession. This has been hampered by the common perception that it is merely a part of computer science. The goal of making software engineering into a profession has spurred a great deal of debate about what it means to be a profession.
To claim to be a profession, software engineering needs to be widely recognized as such. But such things are having undergraduate degrees, licences that are recognized by state governments, their own codes of ethics, and so on are symbols of that recognition.
History
In the mid-1990s, The National Society of Professional Engineers sued in all states in the US to prohibit anyone from using the term software engineer as a noun or field of employment. They won in most (48?) states. Utah does not license professional engineers, so the point was moot there.
In response, the IEEE and ACM began a joint effort called JCESEP in 1993, which evolved into SWECC in 1998 to explore making software engineering into a profession. Both committees used traditional engineering model. The ACM pulled out of SWECC (in 1998?), objecting to its support for the Texas professionalization efforts, of having state licenses for software engineers. The IEEE continued to support making software engineering a branch of traditional engineering.
Ethics
Software engineering ethics is a large field. In some ways it began as an attempt to define bugs as unethical. However that is unrealistic.
Bill Joy argued that "better software" can only enable its privileged end users, make reality more power-pointy as opposed to more humane, and ultimately run away with itself so that "the future doesn't need us." He openly questioned the goals of software engineering in this respect, asking why it isn't trying to be more ethical rather than more efficient.
Lawrence Lessig argued that coding is more like law, in that it expresses a certain social ethic by deciding what to ignore/consider in making detailed decisions.
Some certifying organizations such as the Institute for Certification of Computer Professionals have a formal code of ethics, to which adherence is required as a condition of certification. According to the ICCP, violation of the code can result in revocation of the certificate.
Licensing
Donald Bagart of Texas became the first professional software engineer in the U.S. on September 4, 1998 or October 9, 1998. As of May 2002, Texas had issued 44 professional engineering licenses for software engineers.
The professional movement has been criticized for many reasons.
- Licensed software engineers must learn years of physics and chemistry to pass the exams, which is irrelevant to most software practitioners.
- Many (most?) computer science majors don't earn degrees in engineering schools. These engineers are unqualified to pass the exams.
- In Canada, most people who become professional software engineers actually studied computer or electrical engineering, even though these people already qualified to become professional engineers in their own fields. Thus, professionalism turned into another tool that traditional engineers use to prove that they are better than computer scientists and software engineers.
- http://www.faqs.org/faqs/engineering/pe-eit-exam/ Apr 10, 2002
- http://www.sdmagazine.com/documents/s=746/sdm0003g/0003g.htm
- ACM position on licensing of software engineers, http://www.acm.org/serving/se_policy/
Undergraduate degree
This is currently being defined by the CCSE.
See also
