Building Parpera: Flutter on the front-end 📱

Hello everyone! :wave:

Having worked on the Parpera app for a few months now, I’ve been itching to share some of the technological choices we’ve made and why they’re exciting to us. Today I want to talk to you about Flutter, the backbone of our mobile application.

Flutter is a cross-platform mobile framework created by Google a few years ago which is finally mature enough to deliver quality, production ready applications. It plays a unique angle in solving the cross-platform development problem, essentially providing a custom rendering engine to the underlying platform which allows us to write platform-agnostic code in a single programming language (called Dart). With Flutter we can:

  • Build for iOS and Android at the same time :apple::robot:
  • Work at least twice as fast when compared to native development :zap:
  • Create quality, natural interfaces with control over every pixel on the screen :ok_hand:
  • Write robust, reliable, and secure code that is resistant to bugs :lock:
  • Attract talent with an interest in frontier tech :mortar_board:
  • Tap into a sea of knowledge in an open source community backed by Google :face_with_monocle:

Some of this was covered in our first investor update but I think this is a great forum to provide another level of detail and a better opportunity to answer any questions about the technical implementation of the Parpera platform!

You can read more about Flutter here


Thanks for sharing your thoughts on Flutter @Dave

I know you come from a React Native background which is currently the dominant mobile cross platform development framework, but Flutter has seen a steady rise in adoption over the last few years, with surging levels of popularity recently. What do you think is the reason for this?

Also what are some of the biggest hurdles you’ve had to overcome in picking up Flutter as a seasoned React Native developer?

Have you come across any misconceptions behind the Flutter “hype” that you would like to address?


Yes that’s right @nxsliu , I’ve been developing with React Native since it was released by Facebook in 2015!

I think there are several factors contributing to the rise of Flutter!

If you aren’t familiar with React Native, it is a framework for building native apps using JavaScript and React, React itself being a library that was originally built by Facebook for creating web apps. This enables web developers to transition very easily into app development due to the accessibility of the language and framework.

On the flip side of this, JavaScript is a quirky language that has been through a myriad of iterations since it became popular on the web and it comes with its own set of problems. Some of these problems are mitigated by applying layer upon layer of abstraction over the original language (see Babel, TypeScript), and some by using a mountain of dependencies. Without getting in the specifics, I would argue that it is simply not the right tool for the job.

In contrast, Flutter uses a type safe language (Dart) which was purpose-built by Google for optimal UI development. Dart native features JIT (just in time) compilation which greatly speeds up a developers workflow. With React Native, the JS bundle must be loaded into memory and parsed before it can be executed, but with Dart, the AOT (ahead-of-time) compiler allows apps to start instantly and run smoothly.

I think that React Native developers are frustrated with some of the limitations of this environment and are more willing nowadays to look outside of the JS bubble to other solutions that exist for cross-platform client development. A natural alternative is Flutter.

There haven’t really been any major hurdles to be honest!

This was hardly a hurdle, but there was a natural learning period of a couple of weeks in which I was picking up Dart (which actually felt quite natural coming from JS), and Flutter (conceptual differences between Flutter “widgets” and React Native “components”). This was an overall positive experience, especially with some moments of delight as I realised how quickly some problems could be solved using Flutter which were previously arduous tasks in React Native. Google releases “widget of the week” videos on YouTube showcasing some of these capabilities if you’re interested in seeing some specific examples.

I think the biggest pain points so far have been related to Flutter’s fresher (but growing!) community. I am used to being able to reach out and immediately find a popular, well maintained third-party JS library when creating React Native applications. Given that developers are only more recently starting to solve similar problems in Dart and Flutter, there is less choice out there at the moment for quick-wins from that third-party library perspective. This will be the case less and less as Flutter “grows up”.

Similarly, there is less consensus about the “correct” way to architect a robust, scalable Flutter application. We are seeing developers use architecture patterns and state management solutions that they prefer and that work well on other platforms but the jury is still out on which of these approaches works best. Again, this due to the fact that Flutter is newer and people have had less opportunity to really put some of these approaches to the test in production and make their cases to the community. Often the dust never really settles with these debates either - approaches can vary widely in style and we see cycles of popularity in approach not dissimilar to those seen in the fashion industry.

Ultimately, Flutter is a framework that uses a rendering engine called Skia under the hood to graphically paint everything onto the screen, pixel by pixel. This comes with many benefits, for example confidence that an interface is going to look consistent across the platforms on which it is run and the flexibility to create anything on-screen.

The biggest downside to Flutter’s approach is, in my opinion, that it does not (and cannot) automatically keep up with updates to the underlying framework. For example, if Apple releases a new style of button for its iOS operating system, that same button needs to be replicated as a Flutter widget (either by the Google team, a third-party developer, or ourselves) before it can be utilised in a Flutter application. This leads to a delay in which new platform features can be taken advantage of by users, and can sometimes deliver a sub-optimal experience when compared to the native equivalent.

Due to this, native approaches will always have a slight advantage over Flutter in terms of delivering (what the platform thinks) is a first-class experience. When the realistic net value of this advantage is weighed against all the other benefits that Flutter brings to the table, I still think it’s the obvious choice for small to medium sized teams who want to deliver quality apps, fast.

1 Like

Love the insight Dave, sounds like flutter is the right choice!