Re: User Defined Fields - HELP PLEASE!

From: Mike Sherrill <MSherrillnonono_at_compuserve.com>
Date: Wed, 16 Feb 2005 10:47:56 -0500
Message-ID: <cgp61195eramjnopvke2fdikdap7blsaff_at_4ax.com>


On Wed, 16 Feb 2005 09:05:07 -0600, "news.hp.com" <bunchah_at_yahoo.com> wrote:

>I've got a political (and technical question) I could use a hand with.
>
>I developed (and maintain) an inventory system for a large company. It has
>over 150 regular users, one of which is bucking really hard for 5-10 "User
>Defined Fields". Read that as free-form text fields.

More likely as "free-form text fields with wildly varying semantics".

>Can anyone here present a strong argument as to why doing this as a bad
>idea?

>Management, as usual, doesn't understand that allowing users to put
>any random data they want in these fields is bad in the long run.

Quite often, users will use such fields to sidestep table or application constraints. Whether the constraints are right is a different issue.

For example, the conceptual model might say "Every mechanical drawing has a date on which it was approved, and the earliest possible approval date for a mechanical drawing is '13-Oct-1953'." A user wants to use 05-Jul-1935. You'll probably find today's date in [Date_approved] (because that's a common application-level default), and something like "Actual date approved was 05-Jul-1935" in a free-form text column. This kind of practice renders both columns almost useless.

You really need to know *why* the users want to be able to record "anything at all", and apply a software engineering methodology to manage changes. You can download the latest version of the Capability Maturity Model from SEI (www.sei.org, I think.

It's a real ball of worms, but you need to be able to say "Here's a list of things this database has to do", and "Here's how your change screws up a, b, c, and d."

>It'd be nice if you can point me to reference material where this is spoken
>against (either online or in currently available publications). I need
>something concrete to take with me to a meeting this afternoon.

Might be too late for that. Consider getting really sick and postponing the meeting. <g>

-- 
Mike Sherrill
Information Management Systems
Received on Wed Feb 16 2005 - 16:47:56 CET

Original text of this message