Re: FK -> non PK - bad design?
From: --CELKO-- <71062.1056_at_compuserve.com>
Date: 12 Apr 2003 16:55:37 -0700
Message-ID: <c0d87ec0.0304121555.26d1eb77_at_posting.google.com>
Date: 12 Apr 2003 16:55:37 -0700
Message-ID: <c0d87ec0.0304121555.26d1eb77_at_posting.google.com>
>> So you're arguing that we can know from context (it's in a column
in a normalized database) that it's a scalar, but you won't accept
similar contextual info (it's in a table, which by definition
represents a set) to
know that the set can contain more than one element, even if the set
has a singular name? <<
- From the stuff I see, assuming a normalized database is leap of faith <g>.
- SQL can have both tables and columns with the same name; the syntax
tells the compiler what to look for. When you go to a data
dictionary, you have problems because you are out of context. Maybe
we should have had one name space, but its too late now.
- The other "religous argument" is how to write names (uppercase, lowercase, with underscores, etc.) and we have some research on that.