Architecture

[Alagris]

So far there should obviously be:

• theoretical paper
• C library - depends on paper
• cmd interface & compiler - depends on C library
• online REPL - depends on compiler interface

When it comes to C library i think there are two basic architectures that we could follow:

• there is one data structure and several functions operating on this data structure (possibly in-place to save space). Every more specialised data structure is just a special case of that one general data structure. For example finite state automata are just mealy machines with empty output everywhere.
• there are multiple data structures and some functions take and produce different structures. For instance one function takes 2 finite state machines and combines them into one Mealy machine

For cmd interface, we can take two basic architectures:

• there is one real interface and compilation works just like: first start real, then type ":load some/file.code", then type ":compile", the type ":save some/file.binary" and then ":exit"
• every time you want do perform some operation, you use different command for it. For instance to concatenate two automata you run "compiler --concatenate file1.code file2.code --output file3.txt" and then to compiler you run "compiler --compile file3.txt --output file.binary"

when it comes to online REPL we can experiment with 2 possibilities:

• everything runs on user-side in webAssembly
• mostly runs on server-side and user only sees interface