While I agree with the concerns about ballot stuffing (they needed to have separate castle and keep voting, and some sort of voting oversight that timestamped votes, so that any bloc of votes for one design that happened in a short time would draw suspicion and an investigation - and have that in place without telling anyone)
But...
What was the quote from Big Trouble in Little China about "not getting it"?
It looks to me that the POINT of the contest has gone straight over the heads of many, on all sides of the debate.
The idea was that the contest was to create new variants of the keep and castle that people could use and decorate the way they saw fit, that were actual, usable, designs.
Most of the designs submitted were really nice CUSTOMS, that while they looked good, would function poorly (not to mention would not have broad appeal). The ones that functioned great, looked too much like the problematic 18x18 boxes scaled up. A prime example was that many in both categories (including at least 1 winner) failed to leave spaces large enough for easy use of things like raised garden beds and other large add-ons.
They were looking for something in between, that retained elements of the original, but were new and retained an element of "letting the end customer finish it". The concept of a "town on a castle plot" is nice - until there's two dozen identical towns on one shard, some sitting side by side!
Remember that, if you simply go with a few designs based on only looks or only utility, you WILL have an issue with the "McMansion concept", as the new designs will lose all sense of uniqueness from overuse, or get ignored from not being used. That's why this contest is going to take the better part of a year or more, in different sections, to see what works, and what doesn't. And, the Devs really do need to introduce some sort of proactive means of curtailing the "See what I can do (who cares if you could actually live in it)" designs.
Frankly, what I'd like to see is something akin to the old "Dungeon Geomorphs" mapping aids for tabletop RPGs (look up the concept). Essentially, have the floor plan of a castle or keep mapped out in 9 sections (a 3x3 sectional grid). Then have 5-10 floorplans for each of grid sections, that retained connectors to allow passage from one section to another where they meet. You pick the sections out (and their wall type) as you place them. So, for example, you could choose sections 1-3, 4 and 6 (As a phone keypad layout) to be building elements, 5 being an open courtyard, and 7-9 being more courtyard surrounded by a curtain wall (and you could choose this wall to be a log stockade wall, instead of stone), with doors set in one of the segments of 7, 8 or 9. Another person could have 2, 4, 5, 6, and 8 form a cross shaped building, with 4 corner courtyards (all with door access), with or without exterior walls (and the entrance either through one of the potential doors in 8, or one in 4 or 6). You'd have thousands of potential combos, even before factoring in wall material choices, and the deco added by the owner.