
By considering the detail and scale of the analysed User
requirement, we can take an early “best guess” at the environment and
technologies that will be appropriate for the development, test and eventual use
of the application.
The Developer must provide evidence of the consideration of the metrics of the application.
How big will the database become (bytes)? If this is a multi-phase development, it may be advantageous to select a scalable database that will cope with the eventual full data load.
How many tables? This is a good measure of complexity. More tables result in more forms, pages and reports. It also hints at more complex queries and processes.
How many records in each table? If tables with many records (>50k) are to be included, the database-type, indexes and topography must be considered well.
How much data needs to be returned to the Client? If a file-server topography is adopted, then data is sent to the client for querying or processing. If this data is either frequent or large it may degrade the bandwidth of the network.
How much record recovery? How much extracting of records from tables and recordsets will be required? This is an indicator of the level of expected querying load.
How much aggregation? This also is an indicator the level of expected querying load.
How much number crunching? Should it be performed at the Client or Server?
How many forms/pages?
How many reports?
How complex?
Single User? This will usually be a MS Access client application with a MS Access backend.
Local Network? For light to moderate data loads, this will probably be a MS Access client application with a MS Access backend.
For heavier data loads this will probably be an MS Access application with an Oracle or MS SQL Server backend. MS Access may use Pass-Through Queries (SPT’s) and ADO to communicate with the backend data.
National? European? Global? Once a single source of data is to be available at more than one location, then there may be a preference for a web-based application.
If a backend Oracle or MS SQL Server database were “exposed” to users at other locations, a MS Access Client-Server configuration would still be a good cost/performance solution. However, it would be a little more costlier to roll-out and maintain than a web application
Run-Time? If the cost per user or available MS Access licenses are relevant, then a MS Access run-time version should be considered.
The method of distribution and the software loaded, and its affects on MS Office should be discussed with the Customer.
Total Users? The total number of Users would be the total number of users who will access the application.
Maximum concurrent Users? The maximum number of concurrent Users should be estimated by the User Representative. If this is high, it should be discussed with the customer Record Locking Record Locking strategies must be discussed with the Customer