>Great, that looks like it would work well.

The example I showed was not ordered in alpha. So, it was a little bit difficult to follow. But, if you look at it, you will only find one ID which contains more logs than others. That's the one that got responsible for the initialization. Because I load all my data dictionary in memory, it takes a little bit more time. This is about 1.5 to 2 seconds. But, after that, it is much faster. Basically, I got rid of about 10,000 lines of code in the framework. It does eat more memory. But, it is not a big factor.