I believe they changed the information model a couple of years ago to accommodate the Address Data Management map template. With that template the workflow for adding new addresses is you click once on the access point on the centerline then click on the center of the parcel or on top of the building. The Attribute Assistant then fills in all of the fields for you. It does this by getting the house number by calculating where the first point you clicked is on the centerline using the range for that segment and side of the centerline. It fills in the full road name using a master street name table. It then concatenates the house number with the full street name to get the full address.
I think that they no longer parse out the street name because that is done in the master street name table and any locators you would use with these points would only index the full address. Why build indexes for multiple fields when you have everything in one field?
Using the master street name table reduces errors when you create address points. The above workflow works for new addresses but if you are creating points for existing addresses that method doesn't work because the house number the workflow comes up with may be different than the house number that already exists. I set up the Attribute Assistant to get the house number from my parcel layer that had the house number in a field. I then used a nearest feature method in the Attribute Assistant to get the street name. Sometimes this would get the wrong street if the address point is for a corner lot - and I think that is why they set up the workflow the way they did to eliminate that error.
Scott Oppmann can probably explain this better than I.
Jeff Ward
Summit County, Utah