Well this kills any enthusiasm I had for the project, so much for donating.
HTML5 would have been the best option as the baseline exporter since the code it generates is pretty much open and relatively unobfuscated, which would allow a good codebase to work on to create new exporters. I'm not sure what other exporter could be considered to be a baseline or if you're even considering to develop an open source exporter at all.
This hobbles development of third party exporters significantly, good for your bottom line, not so much for hobbyists. Having third party exporters would have allowed ports to more non-mainstream systems such as the Dingoo and Gamepark linux-based handhelds which wouldn't be able to run interpreted code at a decent speed. I find it highly unlikely that you would find the resources to create exporters for less popular and therefore less profitable systems.
Of course the poll is just an empty gesture, I doubt the results of it would sway your opinion one way or another since once a project starts seriously considering going down a closed path it's basically a done deal.