So far in SI development I have either used AF to house required PI Points for the Input Adapter (search AF, then create an array of tags), or even flirted with PI SDK (which I will now use RDA).
Anyway, I wonder why there is a PISearch object for performing tag searches? Is it in case a SI developer is not using PI SDK or AF SDK; so there is only a dependence on the PI for SI Adapters?
The search criteria you build up using PISearch, is that the equivalent of a GetPoints call in PI SDK? (i.e. same underlying RPC)
Now the political question, how do the development efforts for RDA and the SI Adapters match up? Are they both actively being developed separately? Because it looks like they are using two separate approaches under the hood.