Hi Marc -

Thanks for your help, but I think I wasn't clear on what the actual problem is. I have a mediator that calls a class called Updator. Updator updates a given field. The problem is that I have *multiple updates* to do. The exact number and process of the updates will vary depending on the specific situation.

How do I structure the mediator and Updator so that the 'Updator' process gets called the proper number of times?

TIA -
~ ellen


>Conceptually, I like the ADO model of three independent but friendly classes that are not hyarchically related, but rather exchange themselves as parameters: connections (physical locations, servers etc), commands (talbels and queries) and databases (types: vfp, oracle, etc...)(quoting from memory). It seems to me that you can locate all the aspects of the problem as you specified it into these three classes. Whether this fits Pattern theory I do not know, but, it is my impression that Pattern theory does not help much when it comes to data access.

>>I am working on (struggling thru!) my first real abstract OOP project. I have a 'mediator' that controls 3 processes. Everything was humming along fine until I realized that in the last process (called Updator - where various tables and fields get updated) there are *multiple* items to be updated. These items are not necessarily in the same table, or even in the same *type* of table (could be Access, Oracle etc.) I could need to connect to these tables via VFP, ODBC, ADO etc.
>>
>>I'm not sure design-wise what is the best way to generically handle this situation. At this point, my stumbling block is not the fact that there are different types of data that I am dealing with, but the fact that I have to go thru the update process MULTIPLE times. I'm not sure generically how to set this up.