Well, I'm hoping if we can get Dragon to release his source, we can convince YWG to re-open his SVN and start updating Wood4TT. One important thing to understand about the GPL is that you don't necessarily have to publically publish the source. What you are required to do is, on request, provide the source to anyone you've provided the binaries to. It doesn't even have to be an electronic copy - you can actually provide it on paper after requesting a reasonable fee for printing and postage. If nobody has contacted Dragon to request the source, he's not actually violating the GPL. Ditto YWG.
On the subject of hex editing, it is theoretically possible to do updates in this manner, but it's an extremely difficult and fraught process. Fundamentally, it involves reverse engineering the code through dissasembly (and ideally decompilation) and then reassembling and integrating the changes, updating any addresses and entry points in the raw machine code. Possible, but not really practical - it would probably be easier to rewrite Wood4TT from scratch.
To that end, there's a more compelling alternative; disassembling the same version of both Wood4TT and WoodR4, identifying both loader sections, and working out how Wood4TT manages to load on the DSTT by comparing the two, with the released WoodR4 source as a guide. While I've done some work to this end around this neck of the woods, the truth is I'm just not familiar with ARM assembly (or any other RISC architecture for that matter) and I'm having trouble getting my hands on a good ARM decompiler. (a much rarer tool than dissasemblers, that's for sure)
Regardless, it seems the best course of action is to try and get in touch with Dragon. Does anybody who speaks Japanese want to try?