What is software development?
Software development refers to a set of computer science activities dedicated to the process of creating, designing, deploying and supporting software.
Software itself is the set of instructions or programs that tell a computer what to do. It is independent of hardware and makes computers programmable. There are three basic types:
System software to provide core functions such as operating systems, disk management, utilities, hardware management and other operational necessities.
Programming software to give programmers tools such as text editors, compilers, linkers, debuggers and other tools to create code.
Application software (applications or apps) to help users perform tasks. Office productivity suites, data management software, media players and security programs are examples. Applications also refers to web and mobile applications like those used to shop on Amazon.com, socialize with Facebook or post pictures to Instagram.1
A possible fourth type is embedded software. Embedded systems software is used to control machines and devices not typically considered computers — telecommunications networks, cars, industrial robots and more. These devices, and their software, can be connected as part of the Internet of Things (IoT).2
Software development is primarily conducted by programmers, software engineers and software developers. These roles interact and overlap, and the dynamics between them vary greatly across development departments and communities.
Programmers, or coders, write source code to program computers for specific tasks like merging databases, processing online orders, routing communications, conducting searches or displaying text and graphics. Programmers typically interpret instructions from software developers and engineers and use programming languages like C++ or Java to carry them out.
Software engineers apply engineering principles to build software and systems to solve problems. They use modeling language and other tools to devise solutions that can often be applied to problems in a general way, as opposed to merely solving for a specific instance or client. Software engineering solutions adhere to the scientific method and must work in the real world, as with bridges or elevators. Their responsibility has grown as products have become increasingly more intelligent with the addition of microprocessors, sensors and software. Not only are more products relying on software for market differentiation, but their software development must be coordinated with the product’s mechanical and electrical development work.
Software developers have a less formal role than engineers and can be closely involved with specific project areas — including writing code. At the same time, they drive the overall software development lifecycle — including working across functional teams to transform requirements into features, managing development teams and processes, and conducting software testing and maintenance.3
The work of software development isn’t confined to coders or development teams. Professionals such as scientists, device fabricators and hardware makers also create software code even though they are not primarily software developers. Nor is it confined to traditional information technology industries such as software or semiconductor businesses. In fact, according to the Brookings Institute (link resides outside of ibm.com), those businesses “account for less than half of the companies performing software development.”
An important distinction is custom software development as opposed to commercial software development. Custom software development is the process of designing, creating, deploying and maintaining software for a specific set of users, functions or organizations. In contrast, commercial off-the-shelf software (COTS) is designed for a broad set of requirements, allowing it to be packaged and commercially marketed and distributed.
Steps in the software development process
Developing software typically involves the following steps:
Selecting a methodology to establish a framework in which the steps of software development are applied. It describes an overall work process or roadmap for the project. Methodologies can include Agile development, DevOps, Rapid Application Development (RAD), Scaled Agile Framework (SAFe), Waterfall and others. (See the glossary.)
Gathering requirements to understand and document what is required by users and other stakeholders.
Choosing or building an architecture as the underlying structure within which the software will operate.
Developing a design around solutions to the problems presented by requirements, often involving process models and storyboards.
Building a model with a modeling tool that uses a modeling language like SysML or UML to conduct early validation, prototyping and simulation of the design.
Constructing code in the appropriate programming language. Involves peer and team review to eliminate problems early and produce quality software faster.
Testing with pre-planned scenarios as part of software design and coding — and conducting performance testing to simulate load testing on the applicatio.
Managing configuration and defects to understand all the software artifacts (requirements, design, code, test) and build distinct versions of the software. Establish quality assurance priorities and release criteria to address and track defects.
Deploying the software for use and responding to and resolving user problems.
Migrating data to the new or updated software from existing applications or data sources if necessary.
Managing and measuring the project to maintain quality and delivery over the application lifecycle, and to evaluate the development process with models such as the Capability Maturity Model (CMM).
The steps of the software development process fit into application lifecycle management (ALM). The IBM Engineering Management solution is a superset of ALM that enables the management of parallel mechanical, electrical and software development.
Requirements analysis and specification
Design and development
Maintenance and support
Software development process steps can be grouped into the phases of the lifecycle, but the importance of the lifecycle is that it recycles to enable continuous improvement. For example, user issues that surface in the maintenance and support phase can become requirements at the beginning of the next cycle.
Why is software development important?
Software development is also important because it is pervasive. As IBM vice president and blogger Dibbe Edwards points out: “Software has emerged as a key differentiator in many products — from cars to washing machines to thermostats — with a growing Internet of Things connecting them.”
A few examples:
Soul Machines (link resides outside of ibm.com) uses software to create artificial online advisors that improve customer service and efficiency. The advisors have human faces, expressions and voices that react intelligently, empathetically and efficiently to customer questions and needs. They can answer over 40 percent of customer inquiries without human intervention — and they learn from their interactions to improve over time. Using IBM Watson Assistant to incorporate artificial intelligence (AI) capabilities into the development process, Soul Machines can create and roll out an artificial advisor in about 8 to 12 weeks.
“This is a race,” says Erik Bak-Mikkelsen. “We have to keep up with what’s happening in the market.” Bak-Mikkelsen is head of cloud operations at car2go (link resides outside of ibm.com). He understands that delivering new features and functions to car2go’s ride-sharing apps and vehicles is key to getting and staying ahead. To do so, car2go moved its development operations to a managed-services cloud and adopted a DevOps development model. The result is accelerated development cycles, faster time to market and the capability to scale for future growth.
Working with electrical power lines can be deadly. To stay safe engineers set electrical “lockouts” using physical tags and padlocks to divert power from work locations. French energy company Enedis (link resides outside of ibm.com) worked with IBM Garage for Cloud to develop software that instruments these locks and tags and ties them into a shared network. Tags and locks detect each time they are removed from an engineer’s van and communicate the time and geo-location. As the engineer attaches the locks, their location is recorded on a digital map. All stakeholders share a view of the map to ensure safety, reduce downtime and facilitate repairs. The IBM Cloud Garage collaborative development approach enabled Enedis to develop field-ready prototypes in three months.