On the one hand, we have Fat Guy, who argues that a relatively objective set of criteria should be used in evaluating a restaurant's performance. Recently he wrote:
I wouldn't judge a sommelier's performance based on my idiosyncrasies. If I find a sommelier whose tastes are right on target with my own, I'll go back and re-utilize that sommelier. But it doesn't make me say, "Hey, that's the best damn sommelier in the world because we always agree." If I was giving out the best-sommelier-in-the-world award, that's the last thing I'd look at. Cabrales, it's like with the Supreme Court. Most people assume the Supreme Court justices just sit around and vote whether or not they like the idea of prayer in the schools or whatever. But anybody who studies the subject knows that only the fringe justices do that. The rest of them play by a set of rules called the Constitution. There may be room for discussion about what that set of rules means, but ultimately you've got to acknowledge it as the benchmark.
to support his earlier approach:
Le Bernardin is without a doubt one of the top restaurants in New York, but I rarely eat there because the food just doesn't do it for me. I understand why the food is great, but I don't enjoy eating there. So I don't. I recommend it to people for whom I think it would be an appropriate restaurant, but I don't apply that advice to myself... Ducasse's style of cooking is not one I prefer, but I think his restaurant is the best in New York because it's a legitimate style and he does it better than anybody else does any other equally legitimate style.
The other view, advocated by Cabrales and myself, is that specific flavors that we like or don't like do (and ought to) effect our assessment of the restaurant. I have argued:
If a reviewer doesn't like the way the food at a restaurant tastes, I'd rather read a statement like "Le Bernadin's food was not to my taste, but the fish was of excellent quality and the technique seemed to be exceptional as well" than "Le Bernadin is one of the best restaurants in the city."
For me, the most important characteristic of food is how it tastes. Other factors, such as texture, color, technical wizardry that does not exhibit itself obviously in the flavor, and even Cabrales' humor, can contribute to my enjoyment of a dish, but are not sufficient to make eating it worthwhile. I can objectively report on these other factors, but I can only subjectively report on taste, and this is something that I would hope any person discussing a restaurant would do.
while Cabrales has made the points that:
I have particular ingredients (e.g., chicken, eggs, Brittany lobsters) that I like, but no ingredients (apart from waterchestnuts) that I cannot take in. I generally order what appears most interesting (including in view of a dish's history) and/or most potentially delicious to me. I hold a restaurant accountable for presenting a sufficiently broad and appealing menu that I would find items to order. Thus, I do not adjust for whether the ingredients included in dishes offered are products that I like. If the dish is not subjectively delicious to me, no mitigating factors are available in the manner you suggest.
On the point about fairness, why would a diner take into account a sommelier's recommendation and its appropriateness to the particular diner (recognizing that wine may be have certain "objective" characteristics; however, pairing with the meal and its saucing is rather subjective with respect to subtleties) in evaluating a restaurant, but not the composition and ingredients of the dessert or any other dish?
Okay, now discuss.
Edited to add a recap.