Here at OpenROV we’ve gone round and round on this issue dozens of times over the last three years or so. At this point nobody here that I know of does this with their stock OpenROVs. If you correctly use the adhesive heatshrink to seal the ends of each of the DB-25 wires that comes out of the vehicle, then there is no place for water to get inside the wire.
The drawback, of course, is that once an external wire gets nicked, then there is a path for water, and it can lead to a slow ingress of moisture into the vehicle. If you can identify where the break in the insulation is, then you can patch up over the nick, or splice the wire, or whatever, returning the seal back to its original state.
The main reason why folks don’t want to seal up the wires is that its a tremendous hassle- the channel that the DB-25 wiring passes through does not have that much extra space, and you have to ensure that you don’t have any shorts in the wiring harness before you epoxy things together. Sounds simple, not quite so simple in practice.
As your ROV goes deeper, this seal on the wires gets more critical. For a 100m vehicle, the stock assembly process should work just fine. Right now I’m slowly working on a vehicle to go to the bottom of Tahoe (500m), and it will definitely have “hosing blocks” in the endcaps. Note that the method shown in your link doesn’t completely solve the problem- if the leadwires are stranded, then there is airspace between the individual strands, through which water can leak. So you either need to splice in a short length of solid wire, or you need to tin the small length of stranded wire that you stripped. Taking it one step further still, the small area of tinned wire may not adhere well to the epoxy, so there is still a small leak path involved. Bottom line, there is no easy method that offers complete, absolutely reliable protection from water ingress.
Hope this helps.