With a couple of exceptions, there is one master ASP (active server page) page that is populated based on data from the database selected. This will make a theme update very easy. Within the master page there are Java controls that multiply/behave based on the amount of records displayed. The ASP page communicates to the database through an ODBC connection and I extract the data using SQL commands. This is all child's play.
I add, edit, or delete cubes through a couple of password-protected pages that I designed as well. This gives me the flexibility to add/edit/delete a cube from any internet-connected computer at any time. Nothing extraordinary here either - I've been doing this type of stuff since 2000.
The hardest part, not because it's hard, but because it required so much mental effort, was the design of the site. All the graphics were created in Adobe Photoshop. Writing the code was easy and quick in comparison to this. Creating a whole look and feel that matched what I had in mind definitely took the most time. All in all I came pretty close to my mental concept and I'm pretty happy with the results.
Back - A simple, solid, scalable foundation
The site data is housed on a simple Access database (SQL would've been better, but I didn't want the hassle of dealing with my host to set this up). There's a master table with table names, and a table for each cube category. Within those tables, there's a record for each individual cube. The records are indexed by a master key ID number for quicker retrieval. This part was numbingly easy to setup.