February 8, 2024 by Amanda

Discussion Topic: Code Reviews

The DORA Community Discussions start with a brief introduction talk, and then a discussion with the attendees using the Lean coffee.

Topics discussed:

Suggestions around code review metrics- code review velocity, PR cycle time etc.

Recap from the community discussion:

• The Need for Clear Terminology: The terms "velocity", "code review", and "inspection" are often used inconsistently. Participants agreed on the importance of using specific terms like cycle-time, response-time, or review-rate to improve clarity.
• Metrics: The group emphasized defining metrics based on clear business goals and aligning them with company culture. CHAOSS recommended metrics.
• Potential good metrics: time to merge, iterations, time waiting for a submitter action, how are the reviews distributed across the team, and time waiting for a reviewer action.
• LinkedIn DPH Framework
• Context matters - weekends, holidays, vacations, etc. can affect metrics.
• https://github.com/lowlighter/metrics

Where is the value of code review on a high trust team?

Recap from the community discussion:

• Knowledge sharing across the organization, learning the way others work, make code more homogeneous
• Code review is more valuable when learning is the result

Code Review vs. Automated Testing

Recap from the community discussion:

• Automating Reviews and Testing: using automated testing to streamline review processes, along with reviewing the tests themselves to ensure overall code quality.
• Linters and Auto-Fixing: The value of linters for enforcing style and catching errors was emphasized, along with using auto-fixes within CI pipelines.
• Testing can help save time -- make conversations more valuable: first, check the tests and then ask the more complicated questions
• Adding automation & static analysis into CI pipeline is often called "Continuous Review" (Duvall, 2007). Adding AI into that was getting dubbed as "Continuous Intelligence" ca. 2018 or so (along with "Continuous Compiance")
• Context matters -- age of codebase, team culture, team expertise, industry
• Paul Hammant's blog - The Unit of Work Should Be a Single Commit
• Trust and Test-Driven Development for People

AI replace a human in code review? When, where, and how?

Recap from the community discussion:

• AI vs. Traditional Code Analysis: The potential of AI-powered code reviewing tools was discussed and compared with existing code analysis approaches. It was noted that AI reviewers could go beyond traditional tools, but it's important to see how they function in practice.
• Code Written by AI vs. Code Reviewed by AI: The conversation highlighted the distinction between these two scenarios, with implications for code quality and the role of human developers.
• Be aware of the AI training set
• Human judgement is important

_December 31, 2023 by Amanda

The past year was a fantastic year for the DORA Community of Practice! We've seen wonderful growth, both in terms of the number of members and engagement . We had 30 lean coffee discussions, 15 community connection sessions, and over 170 discussion threads in our Google group.

Growth of the Community

The DORA community has grown exponentially over the past year. We now have over 2,300 (+300% YoY) members from all over the world, representing a wide range of industries, roles, and experience.

2023 Community Lean Coffee Discussion Topics

• Risks for DORA
• Reliability
• ROI and DORA
• A Culture of Continuous Learning
• Scaling DORA at Liberty Mutual
• Deming, DORA, and Operationalism
• DORA in SPACE
• DORA.dev
• Sustainability
• Minimum Viable Continuous Delivery
• Documentation is like Sunshine
• DevOps Roles
• Centralized, Role Based and Team Coordination Models
• Software Delivery Teams
• Team Topologies
• DORA Core
• 2023 DORA Report
• Code Reviews
• 2023 Documentation Findings
• Trunk-based Development

2023 Most Active Threads in the Google Group

• Question: Does Lead Time for Change should contain coding time ?
• Developer Productivity
• Delivery Lead Time: Difference of Criteria?
• What does "Deploy Frequency" really mean?
• DevOps (DORA) Culture Maturity Metrics
• Not sure if this is the place, but... metrics for QA?
• Free/Inexpensive Tool for tracking DORA metrics
• Ranking the DORA capabilities (with some help from GenAI)
• Applying DORA capabilities to safely enable GenAI code generation within the organization
• Question - doesn't deployment time makes the focus on outputs rather than outcome?
• Trunk-based development causes burnout? -- State of DevOps 2023 Discussion
• Follow Up from Forsgren: Software developers can tell you how they feel, but they don't always know what to do

2023 Inaugural Community Summit

The DORA Community Summit was held on October 2, 2023 as a pre-conference event at IT Revolution's DevOps Enterprise Summit.

The slides and [new] videos from the presentations are available now!

A special thank you to our 2023 DORA Guides & Speakers

DORA Guides

• Lisa Crispin
• Steve Fenton
• Denali Lumma
• Betsalel (Saul) Williamson

Speakers

• Apostolis Apostolidis (Toli)
• Scott Aucoin
• Jared Bhatti
• Ben Deany
• Dave Farley
• Brian Finster
• Dr. Nicole Forsgren
• Michelle Irvine
• Nathen Harvey
• Amanda Lewis
• Denali Lumma
• Nikita Namjoshi
• Matthew Skelton
• Dave Stanke
• AJ Wasserman
• John Willis