Caveat and context: I wrote this when things were "VFD" and not "AfD." One must remember that we formerly decided deletions much more along a majority line. For whatever it's worth, I disagreed with a move to "consensus" for deletion, because my general view has been and still is that we should be as sure as we can be that we are not misinforming and therefore ought to be conservative about inclusion. Also note that this speaks of "votes." Even then, I understood a "vote" to be much nearer to the Latin meaning of the word: one expresses a desire, an act of will. Discussion is king, and I hope you see that, throughout, I believe that discussion is the first and last duty of the "voter" on AfD. Discussion should not include fighting, though, and that is what the first principle is about.]
No one has asked me, but here is some advice that's worth everything you pay for it.
You should vote, and you should state the reasons for your vote. You should read the reasons given by other people. If you feel that they are in error, you can explain why, in your vote, their reasoning did not sway you. What you should not do is engage in a dialogue. Most of all do not try to prove your case. Every person who goes to AfD too often will fail this rule. I certainly will. Keep it in mind, though. Remember that the other person is just one vote, same as you, and the strength of your reasoning and his or hers may or may not sway others, but your duty is merely to vote.
This is self-explanatory, until something you agree with in your marrow shows up and loses a vote. At that point, you may feel that real world politics, or snobbery, or feelings of persecution, or gender or personal identity demands that you vote opposite the others and, worse, that you stop the others from doing what they're doing. Remember: these are pixels on copper disks somewhere. This is an encyclopedia, and not real life. Redress the issue by improving the Wikipedia.
The most important thing you can do to be a respected voter on AfD is to be consistent in your criteria. If you can enunciate those criteria, then all the better, but be consistent in them. You have no choice but to obey the deletion criteria, but each one of those has an interpretive element to it. How you understand "original research" or "notability" is going to come into play. However you define these things, be consistent. That means that you should be perfectly happy to delete a non-notable article on a thing you love (or that you wrote) and keep a notable article on something you hate (or was written by your least favorite Wikipedian). Be consistent with your criteria. Explain them, and do not waver. If you vote delete on Sailor Moon because it's too granular, don't then vote keep on Pokemon because people love it, and vice versa.
This is hard. An article appears that has been written by a timid and eager 10 year old, but it's poor. You ought to be judging the article and its encyclopedic qualities, not the person. You can shower the user with Wikilove and offer all the help in the world, but you are going to need to pull the trigger on the article. Similarly, if Dr. Important writes a cruddy article, you're going to have to delete that, too, despite the fact that the author is such a feather in our caps that we daren't risk offense. If DJRadzikul sprays AfD votes with idiocy but writes a good article, you have to also vote to keep that article. The author and the article are separate entities and require separate actions.
This is something that has been violated by Admins in the past. Look, we all have in our heads lists of Wikipedians whom we think are worthless. We all know the ones who announce that they have goals that are contradictory to stated policy. We know the people who vandalize and edit war on articles because they're peevish. That doesn't mean that we try to kill them in pieces by listing every article they write on AfD or, worse, listing their user pages on AfD. People are handled by the Arbitration Committee, not AfD. The only thing that comes out of AfD warring someone is that a pest becomes a troll.
Quite a few people misunderstand or disagree with me, here, but I think it's very important. We needed an article on Charlotte Charke. Someone "wrote" one. That one was, in fact, a paragraph out of context. It turned out much later that the paragraph was, in fact, copied and pasted from another website, from the middle of a biography on another website. Had the article come up for AfD, the correct vote would have been "delete." No one would be saying, "Charlotte Charke is unimportant" or "Charlotte Charke is not notable." What we would be saying is, "This article fails the deletion criteria." What happens if you keep a bad article? We, in effect, still lack coverage of the topic. Arguing that it will eventually get better is, in my opinion, not an argument for keeping. "Eventualism" is an argument for keeping an incomplete article, but never, under any circumstances, an article that violates the deletion guidelines. Thus, vote on the article as it is, not the article as it might one day be.
I think you have more "face" and more prestige if you show yourself willing to reconsider an article after proponents have improved it than you do by insisting that a topic is bad and must evermore be bad. If anyone works on improving an article being considered on AfD, please have the good grace to compliment and thank them for the work. It's easier to say "delete" on a marginal article than to improve it, but, at the same time, it is not the duty of AfD voters to do the improving. Thus, if someone does do the work, even if it ultimately falls short, respect the work, honor the worker, and give a frank reconsideration.
N.b.: I'm told that the following was used by the people who drafted the "notability" criteria. I'm honored, if that's true, even if it was used as a negative example.]
In the spirit of the above, I offer some attempt at enunciating my own definition of the "notable" subject. Because each of the terms in the deletion guidelines are subject to nuance, I have my own set of criteria, and these guide me in assessing notability and the need for separate treatment of subjects.
Obviously, encyclopedias don't tell people what they already know, but is an item so famous that people who do not know the subject will still have heard the term and need information about it?
By this, I don't mean "Have people enjoyed reading it" or "Have numbers heard of it?"
The best of its kind, the first of its kind, or the most successful of its kind is notable within the set of an object.
Finally, though, there is one thing that qualifies a subject for separate treatment. If it is notable but should be best discussed in a master term, it can still be broken out into an individual article if the master article is too large to be manageable. Obviously, that introduces another judgment, but, nevertheless, we need to look at large articles, like Pokemon or BDSM and ask whether we can talk about all the sub-topical matters on these pages. If not, we need to break down the subject. This does not mean automatically going to the atomic level! I.e. just because you can't put every Pokemon card into the Pokemon article, that doesn't mean that each card should have an article. From kingdom, you go to phylum. From BDSM, you go to Bondage gear, then Bondage practices. From Methodism you go to United Methodist, and not your local congregation. Work with the community on logical organization of the material to maintain utility of searches and coherence of presentation.
Anyway, here endeth the sermon.