Yay another person making interpreters on the beam! We are getting quite a collection, from a couple of lisps to a multitude of typed languages to a beam native Lua to others. I love seeing blogs about the actually design though, looking forward to updates!

Yep, recursive decent split lexical parsing steps is very classical, PEGs became more popular about 15 years ago, but they are not taught in schools or anything yet so few know about them, even though they are significantly better designs in my opinion, they can act as a lexer, a parser, both, an AST builder, an AST runner, any mix thereof, all with near the same code, while if implemented properly will run circles around the classical methods in speed while being far easier to read.

Thanks @Qqwy !
I am actually writing this language just to be a very simple and small language ( hence the name ). The aim is not to make anything new or fancy but to show a practical way of writing an interpreter. I found that when I wanted to learn how to build one, I could not find many practical advice and examples. In addition, I wanted to put in practice the theory that I am learning My area of interest is concurrency so I might use it to test some ideas in the future. For the syntax rules, you can see a definition in BNF form of the language in the README. There are also examples of source program in the benchmarks folder.

I know, but PEGs remove the need for the lexical analyzer entirely while making everything far more readable and faster as well.

What kind of reference are you wanting? The Wikipedia article (I’m on phone or is be giving substantially more direct links) talks about the basics well enough, but actual implementations are a bit different. C++'s Spirit PEG library is easily the fastest one in the world and will not be best due to its unique code generation, but it is easily the most powerful form out too with the most abilities except streaming parsing (but you can implement that yourself via a custom iterator or a thread anyway, it can even reverse a custom format back to strings too, or parse binary, or whatever, its PEG engine can parse back and forth between streams of anything to custom data structures of anything). There is one standalone generator it somewhere that can generate PEGs via a custom PEG language that compile to a whole host of languages, forgot it’s name but it’s been around for a long long time and it’s decently fast. There are many blogs about it as well and a few papers. What do you prefer? Or can Google for Parsing Expression Grammers.