May 2018 Summaries

In this series, Bubble's engineering team discusses the inner workings of the cloud-based visual programming language that makes programming accessible to everyone. They explore how storing Bubble apps as JSON data has consequences, particularly with regards to object-oriented programming concepts like interfaces and state. The team designed a second abstraction layer called AppQuery to address these issues, which wraps the original JSONBase classes in a way that provides a compromise between interface and state complexity. This allows for high-level operations to be performed while still working at the right level of abstraction for low-level operations.

With the decreasing cost of technology and the increasing availability of SaaS-based services, launching a venture is becoming cheaper and cheaper, making self-funding a viable option. The founders of Bubble, a visual framework to build web apps, have been bootstrapping their business for 6 years, with a team of 6 and over 150,000 users. They considered the pros and cons of not taking external funding, weighing the benefits against the potential drawbacks. The key question they asked themselves was whether the investors' interests aligned with those of their users. After early conversations with investors, they realized that their interests were not aligned, as investors wanted to focus on on-boarding users quickly, while Bubble's founders prioritized features and performance. However, bootstrapping offers several advantages, including a safer approach to running a business, allowing the company to figure out its revenue model earlier on, reducing the risk profile, and fostering a community-driven culture. Despite some risks, such as the potential for being overwhelmed by competition or working on a lifestyle business that may not be aligned with the founder's vision, bootstrapping can be a valuable option for entrepreneurs.

This is a blog post from Bubble's engineering team discussing their approach to storing and transmitting Abstract Syntax Trees (ASTs), which represent the structure of code in a computer. Unlike traditional programming languages, Bubble doesn't store code but instead stores ASTs as JSON documents. The team faced challenges when trying to load and manipulate these large data structures efficiently, so they invented an abstraction layer called JSONBase, which abstracts away the position in a JSON document from the actual JSON object itself. This allows for faster loading of only needed parts of the application tree into memory without worrying about errors or null references. The team developed three implementations of JSONBase: RuntimeJSON for browser rendering, EditorJSON for the Bubble editor, and ServerJSON for server-side execution. These abstractions enable efficient management of multi-megabyte documents and simplify code development by abstracting away complex issues.