Well I can tell you with total certainty that OQ does NOT backtest 5 years in 1 minute : ) That is quite impressive. Although I said backtesting speed is not a "main concern", if you are indeed able to achieve such a significant increase in speed, that would be interesting for me I would say my main concern is the stability of the system and ease of making modifcations and seeing their results. To give you an example. When i use Tradelink in live trading and I have to make a change, I have to make it in VS2008, build the .dll then restart the Trading program. If I backtest, I have to do the same, then look at the "trades" file in excel and manually calculate my equity curve, drawdowns, ratios etc. Also, it's not easy to pre-load historical data into the strategy. These aren't really digs against tradelink b/c it's open source and all off those features can be implemented by me IF I wanted to but I just really don't have time I'd rather spend on creating and testing strategies. With OQ, the interface is really nice and allowed me to get all of the above done much faster. Also having a decent understanding of C#, it was really easy for me to pick up on how to get things done (people say the API reference is not very thorough, which is true, but for me it was just enough to see the classes, methods and properties and a look through the forums answered most of my questions). I will check out TickZoom. It's just that if it's more like Tradelink than OQ because it's open source, I think I would end up waiting a while for other people to put in some useful functionalities. That's why I am thinking I'll wait until the project gets going and then see if it works for me and if I can contribute.
I've studied NT n OQ but I vote my vote goes to Open Source API that would not be specific to just one Broker as the case is with IB. OQ is better for advanced users than NT and to do anything serious both need to be hacked to convert them to Open Source!