I expect I will be the first, unless someone takes what I've posted here and uses it to create and publish their own before I write up mine (or the original one used for the contest gets published). Still, I've posted a new image file in this format at the end to prove that I've done it.
I've corresponded a bit with Fabrice Bellard. He'd indicated that he simply hadn't found the time to clean up his encoder for publication. To be fair, I wouldn't yet consider mine to be publishable. I'll be very curious to compare them if he ever does -- the neat thing about modern lossy codecs is that there tends to be a lot of latitude for the encoders to make choices.
I'm glad you enjoyed all the detail. I had worried that perhaps there was too much but I like to be thorough.