Jack Syntax Analyzer and Parse Tree Visualizer
Feel free to skip directly to the parse tree visualizer.
Trying to build something from scratch is a great way to learn how it works. More generally, breaking down a complex system to its most basic components then reassembeling from there can be an effective way to understand it. This approach is often referred to as first principles thinking.
A modern computer is an example of a highly complex system. However, its operation is really just a complicated dance of a limited set of basic components. All software ever written is executed by a central processing unit as a series of binary instructions after all. Just ones and zeros. Well, nearly all software… this doesn’t apply to quantum computing, and apparently there are such things as ternary computers?
Setting out to build a computer from scratch is an intimidating task. Where would you even start? One option is to start at the logic gate level. Noam Nisan and Shimon Schocken’s Nand to Tetris course starts from here. The course takes you from this elementary logic level, through virtual hardware programming, machine code assembly, virtual machine translation, and finally high-level language compilation. As you work through each of these levels, the lower level becomes abstracted away.
The course utilizes a simple high-level object-orientated language called Jack, which looks like Java but is much simpiler. Students of the course are eventually tasked with writing a Jack compiler. The syntax analyzer portion of the compiler chews up the Jack code into a series of lexical tokens that are parsed to organize and connect them in a meainingful way. The result of the syntax analyzer is an XML file, which represents the parse tree for a given Jack file.
I set out to write a Jack syntax analyzer in Python, and initially it just functioned as a command-line program. But more interesting was trying to think of a way to programmatically visualize the parse tree for a given Jack class. Since the syntax analyzer returns XML, its output can be used to generate a parse tree visual using a D3.js hierarchy. And the Python syntax analyzer could integrate directly into the browser using PyScript.
The most glaring issue with the current implementation of this syntax analyzer is the lack of error handling. The syntax analyzer does not tell you you have a missing semi-colon for example, it just fails. In the “real world” this would be incredibly inconvienent. For the time being, proofread your Jack code carefully.