...that you keep the information about the physical aspects of the office (within the images) in a seperate table from the data about who currently occupies that space. The physical characteristics are unlikely to change - at least not very often, but the assignment of those offices may have a constant churn.

Along these lines, I'm not sure that you really need to get data entry involved with defining the graphical coordinates. In my estimation, this is a one shot deal that might be better handled as part of the software engineering process. That is, if there is a finite and fixed number of offices to be digitized, then you could have a less friendly way to enter the raw coordinates.

OTOH, the assignment of the personnel to those defined office spaces needs a nice friendly interface because the software staff will want to turn it loose at some point. But since you already have the office space digitized and categorized, all you need in your html front end is an additional mechanism wherein someone can not only view who is currently assigned the space, but also can modify that information (maybe with some security scheme to limit who is allowed modify access).

Of course, if this is a turnkey system where the office space is constantly changing, you probably don't want to hand key the coordinates. Still, you probably want to have a seperate front end for the two tasks, with the coordinate input maybe in a standalone executable and the assignment as part of the normal html interface.

Anyhow, just a thought.