Thursday, July 24, 2008

Spark feedback

Got some spark feedback to day. Well most of the judges thought the ideas were sound but some were concerned with the target market. Probably should have mentioned that, after establishing StockShaping in New Zealand, the effort needed to introduce the service to overseas markets would be a lot less than launching in all markets at the same time. A lot of international companies use small markets as test beds before implementing world wide. If the idea flops, then only a small market is affected.

Carrying on, the main judge saw that the concept had some potential and that we had good timing because of the processing power needed to run the financial models. Now that I think about it, the location that StockShaping would run from wasn't mentioned. The judge might have thought it would run from a consumer's personal computer while we were thinking that it would run from a server farm owned by StockShaping. I didn't think of using GPU's for their GP capabilities like CUDA but I was thinking of using FPGAs for on the fly processing between the database and incoming requests/commands. The FPGA could have a python programme running on it and any database commands will be translated to the correct queries without having to use a soft programme. The benefit of FPGAs would be the ability to update the algorithms while keeping the speed up advantages.

Keeping on topic, the main judge didn't get the impression that we knew how big of a task it would be to create a fully working simulator; especially in the realm of realtime services. The judge also hinted at the possibility of getting postgrad research dollars to pursue this; that ain't a bad idea. Even if the conclusion is that the system is too complex for prime time, I'd still have had the opportunity to pursue the idea.

However, the idea seemed to be exciting and memorable. Good news is that I can still enter in to 10K and Aspire. I'm sure that would make the Aspire lead happy.

No comments: