>Hillmar said:
>>I don't quite understand what is the basic difference between the two approaches, or what is the underlying philosophy of each approach.
>
>Well, I realize now that I did not explain that very well.
>
>The first version has one record per service location with three fields for address codes (only the first of the three is required). The code is the link to a record in the address table.
>
>The second version also has one record per service location, but the existence of more than the required one address is evident only in the address table. The type of address a that a record represents is denoted by a .T. in one of the logical fields.
>
>Does that make more sense?
>
>>For now, let me remind you that you can put the address in a single memo field. Two or more "adress lines" in separate fields seem, to me, more difficult to manage.
>
>Agreed. I will consider doing that for this database.
>
>Thanks,

I am not sure what would be the better design. Probably because I still don't understand! Sorry.

Hilmar.