Building an LSIF Indexer for a Low-Code Platform

In the past 2.5 years of working at Airkit, I’ve been delving deeper and deeper into the fascinating realm of low-code. Amidst this eye-opening yet rather winding expedition, I’ve begun to appreciate the beauty of abstractions. By examining the Airkit’s abstract syntax tree (AST) and my journey of building a Language Server Index Format (LSIF) indexer, we will explore the power of abstractions in the context of a low-code platform.

To understand what low-code is, we must first examine what computer programs and programming languages are. Computer programs are sets of instructions for the computer, and programming languages are abstractions, in the form of formal languages, for these sets of instructions. The amount of abstraction provided to a programming language defines how “high-level” it is.  Effectively, abstractions are what make it possible to write programs in higher-level code without thinking in terms of lower-level code, to write programs in C without thinking in terms of assembly code, and to write assembly code without thinking in terms of logic gates.

At Airkit, we treat low-code as yet another layer of abstraction — a high-level programming language that happens to be visual. As opposed to traditional text-based programming languages that require scanning and parsing to convert the source code into a parsed tree, which later gets converted into AST (abstract syntax tree), visual programming languages have structural editors that directly build the AST.

Leave a Comment