Development Methods Basics

Software development involves several distinct concerns: planning, design, implementation, delivery, operations, and governance. SDLC, agile, DevOps, TDD, and ITIL are not all the same kind of “development method.” Classifying them by scope and level of abstraction makes overlap and gaps easier to see.

What This Section Covers

This section covers project delivery and the operating disciplines needed after release.

1. This page: Learn a taxonomy for the terms and see how the articles fit together
2. Agile vs Waterfall: Compare an iterative development approach with a predictive life cycle, including selection criteria and AI development considerations
3. DevOps, MLOps, and AIOps Overview: Compare operations disciplines for applications, machine learning systems, and IT operations

A Taxonomy for the Terms

Each term addresses a different question and scope. SDLC, for example, describes stages across development, while TDD is a testing practice used during implementation. They are complementary rather than alternatives and can be used in the same project.

Scroll horizontally→

Category	Main question	Examples
Life cycle	Which stages are managed from planning through retirement?	SDLC, AI-DLC, predictive life cycle (waterfall)
Planning and hypothesis validation	What should be built, and how should the hypothesis be tested?	Design Thinking, Lean Startup
Discovery and delivery management	How can problem discovery and delivery run in parallel?	Dual-Track Agile
Team delivery approach or framework	How does a team absorb change and organize its work?	Agile, Scrum, Kanban, XP
Scaling framework	How are plans and dependencies aligned across teams?	SAFe, LeSS, Scrum@Scale
Design or testing practice	How is the quality of code and domain models improved?	TDD, BDD, DDD
Development, model, or operations discipline	How are continuous delivery and production quality maintained?	DevOps, MLOps, ModelOps, AIOps, SRE
IT service management	How are IT-service value, delivery, and improvement managed?	ITIL
Governance	How are IT-related decisions and controls designed?	COBIT

This table scrolls horizontally. Keyboard users can focus the table and use the left and right arrow keys.

This taxonomy is not an industry standard. In particular, sources draw the boundary between ModelOps and MLOps differently, and agile may be described as a mindset, approach, or umbrella term. Compare the target, deliverables, and responsibilities instead of relying on the label alone.

Why Development Methods Matter

In individual projects, the person who builds the system may also operate it. In team development, planning, design, implementation, review, testing, release, monitoring, and incident response are often split across people.

Without an agreed development method, teams often hit these problems:

• The timing for finalizing requirements is unclear
• Every change request interrupts the project
• Quality issues appear only right before release
• Ownership during production incidents is unclear
• AI model and data changes are managed separately from application changes

These approaches and practices turn the concerns into shared answers about who checks what, and when. No single choice addresses every concern; a project usually combines elements from different categories.

Recommended Learning Order

Start by comparing agile and waterfall on a common delivery axis. Then move to DevOps for continuous service improvement, MLOps for operating machine learning systems, and AIOps for applying AI to IT operations.

graph TD
    A["Development and operations concerns"] --> B["Planning and discovery\nWhat to build"]
    A --> C["Development and delivery\nHow to build and deliver"]
    A --> D["Operations and governance\nHow to sustain and control"]
    B --> E["Design Thinking / Lean Startup"]
    C --> F["SDLC / Agile / TDD"]
    D --> G["DevOps / MLOps / AIOps"]
    D --> H["ITIL / COBIT"]

Summary

• Software development terms belong to different categories, including life cycles, discovery, frameworks, practices, operations, and governance
• Methods from different categories are not necessarily alternatives and can be combined in one project
• A useful comparison aligns the target, purpose, deliverables, and responsibilities