YouTube video to accompany these slides: https://youtu.be/iy99k6z4wBg. Making decisions about which open source & SaaS solutions to use isn’t easy. Historically, every company has had its own process for making these decisions, some companies even document them. Now, large companies are converging on a standard process for making technology decisions and recording them. Recently GitHub, Spotify, and eBay have written about Architecture Decision Records (ADRs) and why they’ve valuable, while organizations like the CNCF have created their own Technology Radar to help technical teams streamline the assessment of new technologies. In this presentation, StackShare's Founder & CEO, Yonas Beshawred, unpacks ADRs, technology radars, and how they can be useful for large teams. This talk was originally given at the ELC Summit 2020.
To get a better handle on technology decisions at your company, check out Private StackShare for Teams: http://stackshare.io/private.
12. “An architectural decision record (ADR) is a document
that captures an important architectural decision
made along with its context and consequences.”
- @joelparkerhenderson
24. “Have you made a significant decision that impacts
how engineers write software? Write an ADR!”
- Josef Blake, Spotify Engineering
25. “For example, in 2019 the Creator web engineers
decided to adopt React Hooks. We came to this
decision during one of our bi-weekly web engineering
meetings and after running a small experiment in a
small part of our application.”
36. “Frequently, there are decisions where you want to
make a common choice for multiple occasions — such
as languages and frameworks. For this kind of
important decisions, it’s better to have architecture
governance in place. The two most common
governance tools we found at eBay Classifieds are
building and maintaining a Technology Radar... along
with lightweight RFCs.”
37. “An important aspect of technology radar is the way of
introducing new technologies while also reducing risk.
Every new technology usually should go through two
stages, assess and trial, before being declared as
adopted. The assessment state usually requires
proper analysis, gathering input from your colleagues,
and working on proofs of concept before using it in
production.”
38. ● Objectively assess what's working, and what isn't
● Pollinate innovation across teams & experiment
accordingly
● Balance the risk in your technology portfolio
● Work out what kind of technology organization you
want to be
● Set a path for future success