Low-code and the Future of Software
When taking Harvard’s CS50 class, students are introduced to some of the key concepts of software development including sorting algorithms, object-oriented languages, and the ever-adorable duck debugging. Before any of this, however, they learn about Scratch.
Scratch is an educational tool that teaches young people and novices to create games, videos, and stories using building blocks of commands that can be strung together to create these interactive experiences. I find this to be a fantastic way to introduce both the relationship between a programmer and their program, as well as the idea of code blocks that could be repurposed and combined in different ways. In a sense, it is a no-code development platform that allows the users to direct and command their program (by default, a lanky orange cat) without writing a line of code.
My first real encounter with a low-code development platform was with Microsoft’s PowerApps. I attended a session titled “App in a Day” at a conference and learned to create just that — basic mobile applications in under a day, but I also did not need to know anything about the program running in the background. It felt very empowering! I could create things really quickly, experiment with different ideas and designs, and most importantly, know that it has a good chance of working the way I intended.
After the session, I kept thinking about how the future of software development could be more like building Lego, with the majority of pieces being reused from program to program, and a smaller number of new pieces created specifically for each one. That said, this was only because I had a limited understanding of how professional developers and low-code platforms could co-exist.
In reality, low-code platforms have a lot of potential for improving the users’ lives and expediting market-proofing exercises with quick and dirty MVPs, but they will not fully replace the need for pro-code. A parallel to this is what tools like Canva have done to graphic design or Garage Band to music. While they enable people who are not professionally trained in either discipline and allow musicians and artists to bring their ideas to life faster, they cannot cover the professionals' full gamut of activities.
A day might come where a single platform combines low-code elements with a machine learning algorithm that is so good it could translate what users want to build into code and create the custom elements itself, but I think there is still much growth needed before that is a reality.
