Unit 1: Software and process models

Software is a set of programs plus documentation and data that make it work. Unlike hardware, it does not wear out, but it deteriorates as changes introduce errors, and it is largely custom-built. Categories: system software, application software, engineering/scientific software, embedded software, product-line software, web and mobile apps, and AI software. Software engineering (IEEE) is the application of a systematic, disciplined, quantifiable approach to the development, operation and maintenance of software.

• Failure curves: hardware follows a bathtub curve (early failures, steady period, wear-out); software ideally flattens after early defects are fixed, but each change causes spikes, so the actual curve rises over time.

• Components: programs (source and executable code), data (files, databases, configuration) and documentation (requirements, design, user manuals).
• Reusable components: libraries, frameworks, APIs and services built once and used in many systems — reducing cost and defects.

Phases of the Unified Process: inception, elaboration, construction, transition (and production).

• Advantages: simple, well-documented, easy to manage, suitable for stable requirements. Disadvantages: working software arrives late, changes are costly, risk is high, users see the product only at the end.

• Advantages: clarifies unclear requirements, early user feedback, reduces risk of building the wrong product. Disadvantages: customers may mistake the prototype for the final system; quick-fix code may creep into the product.

• Proposed by Barry Boehm (1986); combines prototyping with the systematic waterfall approach and stresses risk analysis; each loop of the spiral is a phase.

• Advantages: risk-driven, suits large and high-risk projects, changes accommodated. Disadvantages: complex, costly, needs risk-assessment expertise, not for small projects.

• 4GT: tools that let developers state software characteristics at a high level and generate code automatically — report generators, query languages (SQL), screen painters, spreadsheets, code generators and today's low-code platforms.

• Advantages: much shorter development time, higher productivity for business applications. Disadvantages: generated code may be inefficient; tools limited to certain domains; large systems still need design.

• W5HH principle (Boehm): why is the system being developed, what will be done, when, who is responsible, where are they located, how will it be done technically and managerially, and how much of each resource is needed.
• Signs of trouble: unclear requirements, changing scope, unrealistic deadlines, poor communication, lack of sponsor support.

• Measure: a quantitative indication of an attribute (e.g., 12 defects). Metric: a measure relating attributes (e.g., defects per KLOC). Indicator: a metric or combination giving insight for decisions.

• Why measure: to estimate, plan, track progress, assess quality and improve — "you cannot control what you cannot measure".