Tech Radar Update: new blips, old blips
Beep beep. Since the launch of our ITP Tech Radar, there has been a lot of movement on our screen. We’ve added, updated and removed various tools, technologies and platforms in our tech ecosystem. Our Tech Radar needs continuous updates because it reflects the technologiees we love to work with today, and those that are sunsetting on the horizon. To filter out the noise, we’ve made an overview of the biggest changes, just for you. Enjoy!
Experiment: Swift UI
Declarative user interface frameworks are on the rise. React popularised the paradigm, and frameworks like Jetpack Compose, Flutter, and SwiftUI show that the industry is embracing them.
At WWDC ’22, Apple once again announced a series of improvements. SwiftUI 4 will bring more layout possibilities, better navigation, customisable bottom sheets and support for native share sheets. While some things are still hard to achieve with SwiftUI, a lot is possible and often more elegant than with UIKit.
The best we can do is get familiar with SwiftUI and compare our needs to what it has to offer. Know that you can adopt SwiftUI incrementally: from just a specific view to an entire portion of the UI, or a new feature. Do keep in mind that the framework is limited in what it can do on older iOS versions, and isn’t available on iOS 12. But with iOS 16 just around the corner and usage of iOS 12 and iOS 13 continuing to drop, the opportunity to use SwiftUI is growing.
So consider SwiftUI another tool in your tool belt, one that offers rapid prototyping, out-of-the-box support for Dark Mode and Dynamic Type, and ease of adding animations.
Winner: Golang in microservices
Go is an open-source programming language supported by Google that aims to be easy to learn and start with. It is a natively compiled, low-level programming language that keeps things simple. Go treats concurrency as a first-order citizen: it is baked into the language and easy to start with and comprehend. We've seen ourselves relying more on Go over time, especially when efficiency is a quality attribute of the microservices we're building.
Must: YAML CI Pipelines
YAML Based Pipelines for CI/CD are essential to ship reliable software. They consistently shorten development cycle times, proactively detect and solve issues, and deliver more quality and value for our clients.
We don't take a firm stance on which platform you should be working with - GitlabCI, Azure Pipelines, Github Actions, Google Cloud Build, or CircleCI - but running one of these platforms is an absolute must for us when building digital products.
Promising: AWS GreenGrass
We've been exploring Amazon IoT Greengrass as an edge runtime and cloud service to help us build, deploy, and manage device software. The Amazon IoT Greengrass Core enables the local execution of Amazon Lambda code, Docker containers, or native OS processes on edge devices. This can be anything from pre-processing and aggregating data, to inferencing ML models and only sharing back the prediction back to the central cloud service.
The fleet of devices can be managed via the complementary IoT Greengrass Cloud service. We've worked with the platform's second iteration and built some early proofs-of-concept that validate our intended use cases. As we see more Cloud/Edge use cases on the horizon, we believe this technology is worth further investigation and assessment.
ML winner: Streamlit
Streamlit is our go-to prototyping tool to get feedback on the ML models we train. ML helps humans make faster and better decisions, yet they also sometimes make strange predictions, or the capabilities of a model are not always transparent. At In The Pocket, we radically focus on users as we understand the importance of talking to your future users and engaging with them in a fast learning loop.
Streamlit lets us quickly pack a trained model and bring it to a user to gather qualitative feedback. Streamlit turns data scripts into shareable web apps in minutes, in pure Python, the lingua franca for ML Engineers. In our discovery work, we're running multiple Streamlit apps for different clients, so we've promoted it to adopt on the tech radar.
So far for the main highlights on our Tech Radar. Feel free to explore the full tech radar here. We hope you learned something new, and make sure to let us know if you think we should update our beliefs or are missing an emerging trend. We’re happy to further motivate why we make certain choices, and learn together along the way.