Re: Reminder, blatant ad

dawn wrote:
> Jan Hidders wrote:
>

>>[...] Data independence means that the DBMS
>>offers you commands to change the organization at the physical level,
>>but when you issue these commands all your operations at the logical
>>level, i.e., queries and updates, can remain unchanged even if they
>>concern data for which the storage structure was changed.
>>
>>So, to test if you understood this, can you tell me when this might be a
>>useful feature? :-) 

Yes. Or the other way around: cluster data from related logical entities together. Or add indexes, or remove indexes, or add other redundant structures such as materialized views.

> I'm still wondering what the end-user requirements are that would
> prompt such changes. For example, is the reason for data independence
> typically performance?

It makes it easier to respond to changes in access patterns such as the arrival of new complicated queries that require changes to how the data is stored in order to get a reasonable performance. In that respect it gives you more flexibility.

• Jan Hidders