The Brilliance of Data Separation

Just about every new project we have built since 2007 has been based on the data-separation model. In a nutshell, we place all the data tables and the data in one file and just about everything else in another file. And here’s why;

In the previous story, about 90% of the changes were made in the interface file so when we came to upload the changes to the client’s production server we could;
1. Update the data file with some minor changes without disturbing the current users.
2. Log everyone off, replace the interface file with the new copy, log everyone back on. This took less than 5 minutes to integrate nearly 20 hours of programming. The alternative of laboriously cutting and pasting changes from our master files to the production files just doesn’t bear thinking about.