The point of the poll seems to be to determine what people think of the proposal. A poll for that purpose makes sense - lots of people have participated in the discussion, on various subtopics and with concerns that may or may not have been addressed.
What doesn't make sense is a poll to determine the consensus view of whether or not consensus has been reached. You're using an exact metric to measure a nebulous phenomena. Someone made a comparison to voting for Obama or McCain based on who you think is most likely to win, but I think an even more apt comparison can be made to RfA. It's like having a long discussion about an RfA candidate (c.f. Ironholds) and then asking voters "So, did we agree in that discussion or not?"
Either way, people are clearly voting in answer to different questions. Folks who set up straw polls really need to be more careful.
Nathan