Re: RM of [Organizational] Data
Date: Thu, 21 Apr 2005 07:57:44 GMT
Message-ID: <Y1J9e.18576$5F3.14093_at_news-server.bigpond.net.au>
"erk" <eric.kaun_at_gmail.com> wrote in message
news:1114005215.871138.20370_at_z14g2000cwz.googlegroups.com...
> mountain man wrote:
...[big trim]...
>> IMO by implementing these things within the
>> model of the [organisational] data, there may be expected
>> to be a far greater use and utility of the model, because
>> both these items are dynamic and require management
>> (and admin, and wages, etc) alongside every single
>> instance of the data (model) ---- without exception.
>
> I'm unclear how modeling the organizations is different from modeling
> anything else, or allows a simpler model than general data. I
> understand that some parts of it will likely be hierarchical... most of
> the time. Other parts won't.
Modelling organisations is not different from modelling anything else except that its consideration will arise in every instance of use of a dbms, whereas "anything else" may not arise.
Here is an example of an organisation table:
CREATE TABLE [dbo].[organisation] (
[org_frame] [int] NOT NULL (identifier) [org_id] [int] NULL , [org_description] [varchar] (40) NULL , [org_lower] [int] NULL , [org_upper] [int] NULL
)
with data (in this instance from a college management system) such as ...
org_id Description org_lower org_upper ----------- ------------------------------- ----------- ----------- 0 Database Administrator 0 99 1 College Administration 100 299 2 Traineeships and Course Managem 100 299 3 Financial Information 100 400 4 GL Interface 400 499 5 IT Helpdesk 500 599 6 Trainer Portals 600 699 7 Financial - Credit Control 700 799 8 EMS: Exception Management Syste 800 899 9 IT Helpdesk 900 999 10 EIS: Executive Information Syst 1000 1099 12 CEO - Complete Access 100 999 13 Application Administrator 0 1099
The upper and lower values are used to map applications (with integer identifiers (eg) from 1 through to 1099) to each of the respective workgroups
All data is primarily related (in an external sense) to elements of the organisation and the organisation as a whole, and any relational model of the data should include in its scope a look at the external requirements. The current model is totally internal in that the relationships considered are like Codd's 12.
This was fine to get the rdbms engine off the ground (eg; Oracle 1979) but the environment has changed substantially since that time.
AFAIK Date does not mention "organisations" and whereas "business rules" are listed in his index, an entry for "organisation" is conspicously absent.
Yet it is clearly a model of organisational data and cannot be otherwise.
Pete Brown
Falls Creek
Oz
www.mountainman.com.au
Received on Thu Apr 21 2005 - 09:57:44 CEST