Open sourcing is in this is not about the implementation or the CS algorithm. It is more about transparency. I think the idea is that the public should know how tweets are ranked, why tweets show up in timelines and which timelines, what makes tweets popular etc. Imagine Google publishing a document detailing how their system ranks pages aka publishes internal SEO rules officially. I don't know if it is a good idea or not. People with enough resources might be able to game the system (if they don't do it already).