Isn't that exactly what the "animal donation drive" was about?
People basically paid Toady money on the explicit agreement that it would mean animals that they had sponsored would be put into the game?
Legally speaking, that's a contract. That was a business transaction. Money was paid in agreement for a service.
As I read it (and I must admit I missed the original call-to-arms, even if I was in a position to have donated) it was more a "I'm going to put some more animals in, of some kind or other. For a little dosh, I will put in the animals
you suggest.
As and when he goes "I will be developing some enhanced system-specific variation, of some kind of other..." and he accepts money on behalf of one or more people towards the end of producing a 64-bit build (whether or not he directly uses that money to give him the 64-bit system on which to build/test it), he would be as obligated to do as the OP suggested. But I don't think he's done that.
Personally, I don't get the fuss about 64-bit platforms. Yes, you can use more memory at once (registers for addresses of individual memory locations and larger page-sizes now being double the size/square of the original number of addresses, however there are other ways of doing that, and you'll find that when trying to store exactly the same data in memory (except where specifically packed to avoid this) you end up using twice the amount of space, so you
need more memory!
I don't think DF is regularly in danger of breaking the memory limit (typically 3GB, IIRC, at least under certain schemes of memory management). As said, there's other optimisations (code cycles for pathing, for example) that I would rather put down cold hard cash to happen, if I was confident that there was a particular target solution that Toady would accept my sponsorship of (and for which I had the spare cash at hand).