As a DBA about to take up support for some ARS databases I have many questions. Given that developers may tend to focus on the forms and processes how do you ensure that the resulting data structure is 'sensible' ie. normalised to ensure ease of data integruty, lack of data duplication and redundancy, and an extensible model. Also I have read some manuals, and I feel distinctly uneasy about developers/administrators performing actions that will have profound impact on the structure of the database - eg removing a field from a form will generate export of data, drop/recreate of table and import. Not convinced that this sort of action shouldn't be accompanied by other actions to ensure that such changes can be backed out. How can such best practices be enforcred ? Other oncerns - the ability to create indexes for performance through the admin tool (!), also the ease with which it is possible to screw the integrity of the data when inserting new values into 'selection fields' (lookup data for drop down boxes) !!!
Am I being paranoid, or is there anyone who can confirm my worst fears ? Any other gotchas I ought to know and any top tips for dbas ? Thanks mateys.

wot no replies - that really sums up the kind of problem you're going to have. Basically many of these code monkeys don't have a clue about what it takes to build a robust and enduring system that will add value to the organisation by enabling the capture and exploitation of data. The tool does not facilitate any structure or methodology to guide in the delivering of said system...just lets them draw pretty screens. Go on prove I'm wrong....

wot no replies ??? - that really sums up the kind of problem you're going to have. Basically many of these code monkeys don't have a clue about what it takes to build a robust and enduring system that will add value to the organisation by enabling the capture and exploitation of data. The tool does not facilitate any structure or methodology to guide in the delivering of said system...just lets them draw pretty screens. Go on prove I'm wrong....

If you build a companies information system, the first thing you have to do is to modell the system with proven techniques like "ARIS Concept". This includes building ERM /SERM Modells for modelling the data that is used in your information system. Only the lowest Development level in such concepts include details for implementation. All other levels are not linked to any special Development Environment.

So don't blame everything on the ARS. It's only the implementation environment. The planning work has to be done before!!!!

You can build database structures and implement any form of relational datamodell you created with the AR System. Only Thing is, that there are allways 8 System Fields in a Table.
These fields can be perfectly used to implement time-related aspects of your information system.

Data integriety and consistancy has to be enforced by workflow or field restriction (just like in SQL, basicly you build your SQL with the Admin, only in a visual way)

So long,
Andy