Open Source Financial Software for project

[Hank Moody]

For my semester project in my data structures class I need to find some piece of code/software and 1. analyze its functionality and performance 2. identify/improve data structures that hurt its performace 3. possibly expand upon its functionality.

I'd like to find something pertaining to the stock market and it needs to be in C/C++.

Any suggestions? I'm supposed to focus on the performance of the code so I think something related to HFT would be cool.

Thanks for the help

PS. Just to verify I'm only looking for someone to point me towards some code to use, not answers

 

[Tom Maloney]

QuantLib: http://quantlib.org

 

[Hank Moody]

QuantLib: http://quantlib.org

Thanks. I came across that earlier but the size of that project kind of scares me. I wouldn't know where to begin.
I may come back to this but something slightly less complex would be nice.

Also I'm a CS/Applied Math major and I have little knowledge of finance

 

[Daniel Duffy]

What do you mean 'pertaining to stock market'? That's almost everything..

Ideally, writing your own C++ code is best for you, and that should be no problem for CS. Have you already experience of C++?

http://www.wilmott.com/messageview.cfm?catid=44&threadid=86802

e.g. see C++ FD2 for 2-factor PDE.

QL will take you too long to learn. And it's undocumented.

 

[Hank Moody]

Thanks a lot. That looks like it might be doable. I do have C++ experience by the way.

I'm not sure what you mean by writing my own C++ code. My professor wants me to find an open source project and evaluate/modify it rather than writing our own code from scratch.

When I said 'pertaining to stock market' I just meant that I wanted to do something finance related instead of a game or whatever.

 

[marina]

http://www.quickfixengine.org/

 

[Daniel Duffy]

How much time do you have to finish this project?

 

[brotchie]

Alberto Giannetti has written a GPL licensed EOD backtesting framework in C++ called hudson.

http://code.google.com/p/hudson/

I have never used it, but from a cursory read through some of the hudson code it looks like an ideal example of idiomatic C++. Alberto makes good use of STL containers and algorithms and leverages boost where required.

The code base seems small enough (when compared with quantlib) to digest within a few hours. It may also be interesting to tackle a subset of the quantlib codebase; for example, the quantlib finite difference framework. Interesting reading!

 

[Hank Moody]

Alberto Giannetti has written a GPL licensed EOD backtesting framework in C++ called hudson.

http://code.google.com/p/hudson/

I have never used it, but from a cursory read through some of the hudson code it looks like an ideal example of idiomatic C++. Alberto makes good use of STL containers and algorithms and leverages boost where required.

The code base seems small enough (when compared with quantlib) to digest within a few hours. It may also be interesting to tackle a subset of the quantlib codebase; for example, the quantlib finite difference framework. Interesting reading!

Wow thats the exact kind of thing I was looking for. Thanks!

How much time do you have to finish this project?

We have most of the semester so about 3 months or so.