Open source and the flow of information

Toli of Marketcetera makes an interesting point about open source trading system – that the openness of the code base means that end users can make changes without having to show them to anyone else, whether vendors or peers. If you’re developing proprietary trading features, that’s a telling point.

But I think Toli’s point doesn’t quite give us the whole story. How does an end user of an open source system get to the point where they can successfully hack the code ?  Quite often they’ll have to ask questions on a mailing list of peer developers. In the case of Marketcetera those peer developers will be Marketcetera’s own team as well as developers at buy side firms and dealers. Asking questions will reveal information to competing firms…

Building with Caplin

So, let’s imagine you’re building a brand new single dealer channel for your bank with Caplin Trader from Caplin Systems. You picked Caplin because it’s the clear leader in the streaming comet server space, and you want to deliver your etrading app into the browser so it’s zero install on the client desktop. There are other vendors with viable offerings that compete with Caplin’s Liberator as comet servers – Lightstreamer is the clear number two. There are interesting new vendors entering the sector like Kaazing, and there are open source products too. But none of them is as mature or as widely deployed in live production etrading apps as Caplin Liberator.

A server to stream live ticking prices to a browser is just one part of the puzzle of building a web 2.0 etrading app. You need to build the GUI in the browser, and you need an execution system that will handle RFQ/RFS. Unlike the other vendors Caplin provide this in the shape of the Caplin Trader GUI and the Trading Data Source. The GUI is based on the JavaScript widget kit from SoCentric, and the Trading Data Source provides configurable state machines that keep GUI and server side state in lock step during the lifetime of an RFQ ticket.