RE: Interesting question and debate on ddl tables with datefieldsthatwill not always have a value

A proper system would have a rehire date, as well as reason codes to
describe the termination. CYA, you know?

Paul Nelson
Office 512-392-2577
Cell 708-670-6978
nelsonp@xxxxxxxxxxxxx


-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of
lgoodbar@xxxxxxxxxxxxxx
Sent: Wednesday, October 15, 2008 10:14 AM
To: midrange-l@xxxxxxxxxxxx
Subject: RE: Interesting question and debate on ddl tables with
datefieldsthatwill not always have a value

A good employee tracking system should have an employment status flag.
"Obviously", one would check that flag rather than a null termination
date.

More relevant: what if the employee is rehired? Should the termination
date be reset to null? The employee was actually terminated (or perhaps
laid off) in the past, and setting a null value loses historical data. A
decent system will have an employment history table; but most I've seen
have standalone hire and term date fields as well.

Loyd Goodbar
Business Systems
BorgWarner Shared Services
662-473-5713

-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Joe Pluta
Sent: Tuesday, October 14, 2008 5:16 PM
To: Midrange Systems Technical Discussion
Subject: Re: Interesting question and debate on ddl tables with date
fieldsthatwill not always have a value

Walden H. Leverich wrote:

Use nulls, that's what they're for!

I'm not picking on you specifically, Walden, because many people have
contributed to this thread, but yours is the most on-point statement.

Anyway, not everyone agrees with your blanket assessment of nulls -
including folks like Chris Date and even Dr. Codd himself. A null in a
value (in this case, the termination date) could mean lots of things: it

could mean that the employee is still working for us (meaning there is
no termination date). But it could just as easily mean we don't know
what the termination date was, that the termination date we have is
invalid, or that the termination date hasn't been entered yet.

No, just about every database scientist whom I've read would insist on a

flag that indicates the status of the employee rather than relying on
the NULL value to give you any more information than the fact that you
don't have a termination date.

I won't belabor the point. I just want to make sure that the notion of
databases that is currently in vogue today often differs dramatically
from the folks who actually came up with the concept. That's not to say

that there is One True Database Design; Codd and Date actually disagreed

on nulls. But neither one agreed with the SQL approach.

So anyway, while the general assessment on this list of the use of nulls

is probably the most common one, that doesn't make it right. And it
certainly doesn't mean that RPG programmers who don't like nulls are
some sort of Luddites; indeed, it just means that they choose not to use

a specific technique - one that has arguably been misused or at least
overused by an entire generation of SQL programmers.

Joe