Address Locator efficiency

[LakeGIS]

I have an Address point feature class and a Place Name point feature class.  Each has an address locator built for it and participates in a composite address locator along with a street center line locator and several other locators.

My question is, is it more efficient this way - to have a separate point feature class for the Place Name data - or should the Place Name data be just a table that is used as an alias table inside the initial Address point locator?  It works as is, but I'm wondering about efficiency and best practices.

[JoeBorgione]

Define efficient.

Personally, I see place names and address points as two very different feature classes, even if you have duplicate points in the two.  I keep mine separate, and in a composite with centerlines dispatch 2000 +- 911 calls a day.