diff --git a/README.md b/README.md
index 32f22798..15175253 100644
--- a/README.md
+++ b/README.md
@@ -44,6 +44,10 @@ I am a fan of this design pattern over parser generators for the following reaso
 - Debugging is much easier with hand-written code
 - It is far easier to extend and make dialect-specific extensions compared to using a parser generator
 
+## Supporting custom SQL dialects
+
+This is a work in progress but I started some notes on [writing a custom SQL parser](docs/custom_sql_parser.md).
+
 ## Contributing
 
 Contributors are welcome! Please see the [current issues](https://github.com/andygrove/sqlparser-rs/issues) and feel free to file more!
diff --git a/docs/custom_sql_parser.md b/docs/custom_sql_parser.md
index a660978c..9a425181 100644
--- a/docs/custom_sql_parser.md
+++ b/docs/custom_sql_parser.md
@@ -4,3 +4,5 @@ I have explored many different ways of building this library to make it easy to
 
 The concept is simply to write a new parser that delegates to the ANSI parser so that as much as possible of the core functionality can be re-used.
 
+For an example of this, see the [DataFusion](https://github.com/datafusion-rs/datafusion) project.
+