The OCaml typedtree (post-typing AST) is rather intricate and contains details about the type inference process that you might not want to have to know about.<div><br></div><div>If you don&#39;t need the full width of the OCaml language features, I would recommend starting from a smaller codebase. Rewriting a parser for a subset of the OCaml language is easy, type inference difficulty depends on the features you have in your language, but there are numerous smaller projects you could reuse.</div>

<div>See for example [MinCaml], a minimalistic compiler for teaching purpose, or the [poly] language from the Programming Language Zoo.</div><div><br></div><div> [MinCaml] <a href="http://min-caml.sourceforge.net/index-e.html">http://min-caml.sourceforge.net/index-e.html</a></div>

<div> [poly] <a href="http://andrej.com/plzoo/html/poly.html">http://andrej.com/plzoo/html/poly.html</a></div><meta http-equiv="content-type" content="text/html; charset=utf-8"><meta http-equiv="content-type" content="text/html; charset=utf-8">