xiongyf04:

I am writing a compiler using Haskell. After the compiler parses program, the program is stored into an syntax tree stucture defined blew:

...... data Exp = Plus Exp Term | Minus Exp Term | Term Term deriving Show

data Term = Times Term Factor | Div Term Factor | Factor Factor deriving Show ......

This is just part of the definition. The full tree contains much more definition than this. Now I want to adjust the syntax-tree. However, I don't need to adjust all the data types, but a small subset of the syntax tree. e.g. I might adjust the "Times" data like the following, but not modify the rest of the syntax tree: transformTerm (Times t f) = Times t (FactorInt 100)

However, in order to apply the modification like this, I have to write a series of function to traverse the tree until I get to the Term data type. e.g. I have to define: transformExp (Plus e t) = Plus (transformExp e) (transformTerm t) transformExp (Minus e t) = Minus (transformExp e)(transformTerm t) transformTerm (Term t) = ...

This is tedious and error-prone. I want to know if there some means in Haskell to write a single "generic" function to traverse the syntax tree and only stop on the "Term" data type. Can anyone tell me something about it? Thanks a lot.

The "Scrap Your Boilerplate" series covers this, as does several other works on generics in Haskell. Here's a good place to start: http://www.cs.vu.nl/boilerplate/ also, http://www.informatik.uni-bonn.de/~loeh/SYB0.html http://www.informatik.uni-bonn.de/~loeh/SYB1.html -- Don