Re: RM of [Organizational] Data

"Kenneth Downs" <knode.wants.this_at_see.sigblock> wrote in message news:mri7j2-sms.ln1_at_pluto.downsfam.net...
> mountain man wrote:
>
>> The issue of the ownership of data is possibly worth exploring.
>> Here we are restricting consideration to data held in a database.
>>
>> Using the following list of roles associated with any database
>>> ==================================
>>> DATABASE SYSTEMS ROLE-TYPES
>>> ==================================
>>>
>>> --------------- Internal to the organisation:
>>> I01 - business owner(s)
>>> I02 - business executives and managers
>>> I03 - general organisation work-groups/end-users
>>> I04 - DBA (for SQL-DBMS)
>>> I05 - IT manager
>>> I06 - internal programmers
>>> I07 - specialised development teams
>>> I08 - Operations & help desk personnel
>>
>>> --------------- External to the organisation:
>>> E01 - contractors and consultants (in any roles defined above)
>>> E02 - contract programmers (or software house(s))
>>> E03 - consultants and suppliers (of selected RDBMS software)
>>> E04 - consultants and suppliers (of other software & hardware)
>>> E05 - business, management and financial consultants
>>> E06 - consultants in Models of Data
>>
>>
>> All other roles apart from I01 (buiness owner(s)) are what
>> might be termed custodians (of varying degrees) of the data,
>> whereas the actual ownership of the data resolves to the
>> owner of the organisation. Any diasagreements here?
>>
>
> nope.
>
>>
>> Consequently, implicit in any model of the data should be
>> the understanding that the data ultimately belongs to the
>> business owner.
>
> ...not sure where you are going...

Whatever theory exists of the data, and the data model, and the database management system, that theory should be aware of the implication that the data relates to an organisation, and has ownership by that organisation.

Codd's 12 rules, for example, do not reveal any expression of this implication, and are restricted to a focus that is purely technical wrt the dbms.

>> Thus, implied in the phrase "RM of the data" is the
>> expanded form "RM of organisational data", because
>> data is always associated with an organisation (treating
>> an individual as a minimal organisation) without
>> exception.
>
> ...ok, I think I see the point, maybe. As far as a database model goes,
> it
> need not be concerned per se with data as it may be handled by other
> owners
> after exports or before imports. It would define boundaries which are
> crossed by import/export, but is free to enforce its own design ideas for
> a
> particular owner without considering the design ideas of other owners.

I was considering whether or not the collection of all owners of organisations have any set of common requirements, either in respect of external contraints, or internal constraints to their organisations.

From the perspective of the theory, if certain specific data related commonalities can be shown to exist across the spectrum of all organisations, then one might expect a theory (of data) to address such issues.

> This would be important because it means the RM never needs to "include
> XML" (must put that term in quotes because it is utterly meaningless) just
> because other db owners may use it. Rather, a particular db
> implementation
> must only be able to translate that format.

I have the same "philosophical objection" to "XML", I think. ;-)

Pete Brown
Falls Creek
Oz
www.mountainman.com.au Received on Sun Apr 17 2005 - 05:06:41 CEST