First off, remember that I offered these as suggestions of ways to attack the problem, so this was just my version of a brainstorm idea, and second, I have never used any GIS functions, I don't even know what you're refering to, but I think I can get and idea from your discriptions.
It sounds like you're on the right track. If you can get the photo size to match the size of the footprint of the building, or vice versa, then, measure the size of one side of the footprint in the real world, (get the width of one of the buildings) use the tapemeasure to rescale based on the width of that building. Then just use it to measure the heights of other buildings in the photo.
I'm pretty sure that to do this accurately, you're going to have to use photomatch, rather than just painting a photo on as texture, because photomatch takes into account the perspective of the photo, and thus, measurements are related to the "vanishing point" of the photo, regular imported photos (I don't think) will do this.
This all sounds like a fun project. Also, if you can make this work, you also might consider posting a tutorial in the turorial section.
GT