I know this post is about 30 months late, but Kristen's question in this comment prompted me to finally explain it. When the blog first started, part of what we were doing was using it as a scratchpad for our Zagat reviews. For those that only read the guide, and don't submit votes, the voting guide is as follows:

Ratings
Rate each establishment on the quality of its Food, Appeal, Rooms, etc. on a scale of 0 to 3:

• 3 - excellent
  
• 2 - very good
• 1 - good
  
• 0 - fair-poor

Personally, I actually like this rating system, better than a typical 4 star (and then, if you are the Chronicle, cheat and use half stars), or a more dreadful 1-10 rating. We just don't believe a rating scale with that much granularity can be useful. You will note that it's more of a power law distribution rather than linear scale. The large majority of restaurants will end up being rated 0 or 1, with some as 2, and very precious few as a 3. I probably end up being more strict in my ratings than your average balloter, with 3 given out very miserly, for experiences that were truly transcendental.

We try to make the food rating as objective as possible, meaning that we don't factor in cost or cuisine (there is no 'it's a 3 for diner food'). While our primary impressions simply come from how the food smells and tastes, we also factor in plating and presentation (particularly for the 2 and 3s). We are probably a little biased towards complex but balanced dishes, and are a bit fanboyish for the molecular gastronomy cooking techniques.

Decor is probably the least important rating to us. As long as it passes Bourdain's bathroom test (and sometimes when it doesn't, *cough* Tu Lan *cough*), we will eat there if the food is good. Interestingly enough, this is one that Alexis and I will disagree on most often. Alexis tends to judge the decor along her own personal aesthetic which prefers clean and modern lines. I will give restaurants more leeway if the decor is focused, consistent and purposeful regardless of specific style and if they are trying to create a mood to go along with the dining experience. We are both suckers for open kitchens though.

For service, while we like and appreciate the friendly, chummy waitperson, we tend to get really blown away by the Thomas Keller school of thought: ultra professional with a preternatural ability to anticipate your needs without interrupting the dining experience. We also require servers to be knowledgeable about the food. One of my pet peeves that I've infected Alexis with is if the waiters write down the order. I really feel like if they were familiar with the menu, they shouldn't need to write it down. One test that Alexis and I inadvertently stumbled on is the sparkling water test. Alexis prefers tap, while I like sparkling, so when the waitstaff refills the water, they need to be able to distinguish. We find that it ends up being a pretty good test of the front of the house, in terms of training and teamwork, or foresight to come up with an unobtrusive way to indicate the difference.

Related Link:
Watching the Watchmen: Michelin Stars