Claire, what you discribed is basically what I've done in the field, but without the iterations.
It works just fine, from the photographers' point of view. It works very poorly from the aquarist's/ichthyologist's perspective. The problem with it is that most fish deport poorly after being relocated. We typically want to get pictures of the fish deporting well, frequently try to capture courtship, mating, or parental behavior. On the whole, its best to find a way to shoot the fish in their home tank.
One of the problems I've had in the field has been that my freshly-captured specimens have sat on the bottom, fins clamped and sulking, or have displayed stress coloration, not the normal. Photographers who don't know the animals are insensitive to these considerations.
The photographic problems of shooting fish in aquaria are pretty trivial. Exposure, primarily, controlling reflections, secondarily. The hard problems have to do with inducing the fish to pose/perform where its convenient to shoot them. The real masters of the art are much better at this than I and my mob of peers, all tied for third-best. I've found it very useful to study the work of people like H. J. Richter and A. van den Nieuwenhuizen to try to puzzle out what they did to get the fishes to perform where desired.
The iterative procedure you suggested strikes me as too much work. FWIW, when I was setting up, it took less than one roll of film to solve the biggest problem, exposure. I just took one shot at 1:4, 1:2, and 1:1 at every marked aperture on my little 55/3.5 MicroNikkor from f/3.5 to f/32. That told me which aperture to use at the test magnifications with my standard flash setup; interpolation works well for intermediate magnifications. Why screw around more than necessary?
You're absolutely right that clean glass, inside and out, is a prerequisite for good fish pictures.
FWIW, a very effective way to photograph fishes in the field was invented, AFAIK, by Jack Randall. If you want to search for him, he's John A. and AFAIK he's still at the Bishop Museum in Hawaii. His trick is to kill the fish, spread its fins and fix them in position by painting them with formalin. Then he lays the fish down on its side on a bed of nails submerged in clear water in a box whose interior is painted black. Shoots with the camera pointed straight down, uses flash illumination, the flashes' axes at 45 degrees to the water's surface. He's produced some superb pictures this way. But it has to be done quickly after killing the fish because some fishes' colors fade after the fish is killed. Friends of mine at Smithsonian Tropical Research Institute came up with a variation, vertical bed of nails in an aquarium.
I thought I was clear that I wasn't presenting my practice as the best practice, although it is. Back when, I surveyed all of the practitioners I could find. We all shoot fish essentially the same way. There's only one way to do it that works consistently well. Its a specialized activity, non-practitioners don't have much of a clue. Why should they?
One other point. Years ago I sent a portfolio to Animals, Animals, eventually dropped by when in NYC to discuss what we might do for each other. They were very positive about my work's technical quality, dubious about commercial prospects. This because at the time I was mainly shooting fish against plain backgrounds and they believed the market wanted "natural" shots of fish against nice vegetation. Well, in nature that situation is very rare.
Cheers,
Dan