Velocity is Not a Metric
We once chased “velocity” like it was a stock ticker. Then came the sprints with 40 story points and little actual progress. 2014 taught me a key lesson: velocity is not productivity. Real metric...
We once chased “velocity” like it was a stock ticker. Then came the sprints with 40 story points and little actual progress. 2014 taught me a key lesson: velocity is not productivity. Real metric...
In the ThoughtWorks Radar (Mar 2014), the rise of event-based tracking was clear. Everyone wanted to log everything — page views, API hits, job states. But I learned the hard way: tracking without...
Everyone’s talking about lambdas in Java 8, and rightly so. But here’s what I’m more excited about: Streams API: expressive, readable, and parallelizable Optional: finally formalizing null ha...
There’s a hidden feedback loop in every engineering org: how you slice your systems affects how people collaborate. I saw this play out firsthand as we decoupled a monolith into more modular servi...
In early 2014, I began a focused push to make all critical services idempotent by design. Up until then, some retry mechanisms were “best effort.” But distributed systems don’t forgive ambiguity. ...
2013 was a growth year—not just in traffic, but in architecture maturity. Here’s a candid list of calls I made that aged well (and some that didn’t): ✅ Right: Decoupling ingestion and processin...
Redis was (and is) amazing. In-memory speed, rich data types, pub-sub — a Swiss army knife for backend engineers. But I learned in 2013 that abusing Redis as a database is a recipe for tears. Wha...
The ThoughtWorks Technology Radar (Nov 2013) listed Web Components as something to assess. That felt exactly right. Shadow DOM, templates, and custom elements promised true encapsulation—without r...
2013 was a year of polishing my toolchain. I embraced minimal, fast tools that stayed out of the way. Top picks: iTerm2 with oh-my-zsh Alfred App for clipboard, launchers, and workflows Das...
As systems grow, so does the probability of failure. In 2013, I was deep into building pipelines that needed to survive crashes, partial state, and retries. We stopped aiming for “no failure” and ...