I toyed with this idea a little when the suggestion came from a relative.
My thoughts were,
first, it has to be a boat with a large hull section that is basically flat already, that would easily be accessible for viewing underwater, and yet not be a heavy traffic area in the boat. And flat so that it would not freak out the water flow on the hull surface.
two, there should be a way to cover the Lexan when not viewing, to protect it.
Three, the hole in the plywood can't compromise necessary strength for intended hull performance. Stringers need placement to make up for the absence of skin strength that would have been provided by plywood in that section, just like headers in a house frame replace some strength in doorway and window cutouts where one stud or more is missing. Don't rely on the lexan to replace that strength!
Four, the installation must not be compromised by hydrodynamic forces while underway, such as being pushed up into the boat, or sucked out. If there's a rocker, watch out for bernoulli sucking it right out of the boat. The installation has to distribute loads well on the edges of the lexan so it won't fail at a screw hole or such.
Five, you must consider the different expansion and contraction properties of the lexan compared to plywood stabilized by fiberglass. It will expand more and shrink more with temperature changes. So some give has to exist horizontally in the installation, like floating panels in a wood door, and a flexible sealant used that can accomodate expansion and yet grip during contraction, and yet also be easily removeable for any necessary maintenance of the cutout frame or lexan, but also be utterly watertight. bolting to a more permanent gasket, and then applying flexible sealant in the edges might do it.
Six, fully research these points and lexan properties before proceeding since you must engineer the mounting mechanism and the sealant strategy to accomodate these issues.
Use lots of floatation foam to make recovery easier during R&D tests!
And console yourself with the option of patching it with plywood and glass and cleat/butt blocks, as if it was a repair, should the experiment require more attention than you planned on giving it.
Seven, find out how others have suceeded or failed, and all the other stuff that in my ignorance and desire to go to bed now, that I have left out...
I like the idea, but I have a sailboat to finish first ... and 30 other projects ...
ks