This isn't just a hypothetical, either. This was the problem I ran into with the Clojure IDE that I was writing[1]. I had to drop down to C to write the lexer/p

This isn't just a hypothetical, either. This was the problem I ran into with the Clojure IDE that I was writing[1]. I had to drop down to C to write the lexer/parser and to store the AST. Because of this design decision, traversing, searching, and manipulating this AST was much less pleasant and more error-prone than it was in the initial Objective-C prototype. This is where Julia will shine for me. It will let me "drop down" to more efficient mechanisms when I need to write efficient algorithms, allowing me to use the same tools at all levels. This is actually my new plan for my IDE, I hope to embed Julia in it and rewrite most of the app in Julia. This should also make it dead-simple to let users extend the app by scripting it in Julia, while still having first-class access to the same internal ASTs used by the app itself. [1]: https://github.com/sdegutis/Leviathan https://github.com/sdegutis/Leviathan -- but don't use it, it's in a state of flux