The Financial Information eXchange (FIX) Protocol in F#

There is a myth that F# is only useful for mathematical and scientific applications. This site describes using F# to model FIX messages, which is neither mathematical or scientific. Why should you care? what does F# give you as a developer that C# or Java does not? FsFIX is going to show that

  1. with F# some runtime errors become compile time errors.

  2. property based testing is supercharged compared to unit testing1 and will catch some of the runtime errors that remain.

  3. the code base is smaller and has less boilerplate.

FIX code generation is the first stage of project to build a fully fledged FIX engine using F#. FsFIX generates F# types to represent FIX messages, and also generates functions to read and write these types to and from byte arrays. The source for FsFIX can be found here. At the time of writing FsFIX can be used to connect to applications using other FIX 4.4 engines and feed them with random but valid FIX messages, in order to test the parsing machinery of both FIX frameworks. Errors were found in QuickFIXN (an open-source C# FIX engine) and FsFIX (found and fixed) in this manner, as was an issue with the FIX 4.4 spec.

The unreasonable effectiveness of Algebraic Data Types

Property testing FsFIX message read-write code

Generating F# from FIX4.4.xml (and generating FIX xml from F#)


Performance compared to QuickFixN and QuickFixJ

Building and Running FsFIX projects

1 unit tests are still useful, property testing is not always appropriate.

About the author

My other site, an F# key-value store: https://fredisnet.org/

Interested in hiring me or asking me questions about FsFIX? my email address is my GitHub user-name followed by ‘@hotmail.com’