The Ten Commandments of Practical Optimization
1. Not everything needs to be stated explicitly in the rules; some things just are.
A human doesn't have a hundred and fifty-seven arms, even though the rules don't explicitly say that he doesn't. A character doesn't continue running around after he dies, even though the rules don't explicitly list any negative effects for death. If the designers spelled out every single thing explicitly...even the glaringly obvious...the core rulebooks would be larger than the Encyclopedia Brittannica, and would likely cost as much as a Ferrari.
2. "The rules don't say I can't!" is not practical optimization.
The second commandment is like unto the first. There are many things that the rules don't explicitly say you can't do. The rules don't explicitly say you can't do the "I'm a Little Teapot" dance and instantly heal back to full starting hit points as a result. The rules don't explicitly say your first level character can't have a titanium-reinforced skeleton and cybernetic weaponry. This is because the rules are structured in such a way as to tell you what you can do--not what you can't. An underlying assumption is that, apart from common-sense actions which anyone can perform, the system will tell you if a given character has a given ability.
3. RAW is a myth.
This is one of the dirty little secrets of the board. The Most Holy RAW is invoked continuously by those who want to give their arguments the veneer of officiality. The problem is, RAW is generally applied not as "The Rules as Written," but rather as "The Rules As I Interpret Them And You Can't Prove I'm Wrong, Nyeah." The RAITAYCPIWN. Not quite as catchy an acronym, granted, but that's what it boils down to. This game cannot be played without interpretation and the judicious application of common sense. Try to play the game strictly and exclusively by the rules as written, and you have an unplayable game. Using "RAW" as a defense is similarly meaningless--particularly when your defense rests on interpretation. If you're going to claim that your build is RAW, you'd better be able to make sure that the rules specifically uphold your claim...not simply that they're sort of vague and COULD be interpreted in such a way as to not FORBID your claim.
This becomes particularly important when your claim is especially controversial. Yes, builds should adhere to the rules as written. Yes, any exceptions to that should be noted. But the RAW as some sort of entity unto itself, capable of rendering a build immune to criticism, is not a useful construction, and causes more problems than it solves.
4. Common sense is not a bad thing.
The rules were designed to be read with common sense. Yes, common sense will vary from person to person, but there has to be some basic level at which we agree on core assumptions, or the game is meaningless. If we have one interpretation of the rules where two levels of a prestige class give you infinite caster level, and another interpretation where two levels of that same prestige class give you two caster levels, then common sense tells us that the latter interpretation is the correct one. If a character reaches negative ten hit points and dies, common sense tells us that he doesn't spring back to his feet and continue fighting unimpeded.
5. Intent matters.
I know, I know..."Blasphemy! No man may know the intent of the Most Holy Designers!" Except that, in some cases, we can. In some cases, the intent is glaringly, painfully obvious. In other cases, the intent has been clarified by various WotC sources, such as CustServ. It makes sense to take these sources at their word, people. They work with the folks who design the game, they have access to them. If a conflict comes up, then it can be resolved, but I can't help but notice that for all the talk about how CustServ never gives the same answer twice, they've been remarkably consistent of late. It's one thing to say "This rule is vaguely worded, and we don't know the intent." It's another thing to say, "The rule is vaguely worded, and therefore I can ignore the intent." The first is sensible caution; the second is rules lawyering. When an ambiguity has been clarified, that should be the end of it.
6. Mistakes happen.
Everybody's human. You're human; I'm human; the folks at WotC are human. Sometimes, humans make mistakes. That shouldn't be seen as an opportunity to break the game. Take the Vigilante from Complete Adventurer, for instance. Anyone out there seriously believe that his rather abrupt jump from 1 third level spell at level 6 to 20 at level 7 is NOT a mistake? There are two ways to deal with a mistake like this: a sensible way, and a silly way. The sensible way: "Hmm. There's a column for fourth level spells with no numbers in it, and a column for third level with numbers that can't be right in it. Clearly, this was a typesetting error, and the second digit in the third level spells column is supposed to be in the fourth level spells column." The silly way: "Rules are rules! The rulebook says 20 third level spells at seventh level! If you do it any other way, you're houseruling! I'm gonna make some GREAT builds based on this rule!" Basing a build on an obvious mistake isn't optimizing; it's silly.
7. Simple Is Good.
There are a LOT of WotC sourcebooks out there. I did a rough estimate on the value of my collection just of hardcover rulebooks; it cost more than my car. Not everyone has that kind of cash to spend on this hobby. Not only that--a lot of people simply don't have the time to commit several thousand pages of rules, hundreds upon hundreds of prestige classes, and thousands of feats to memory. So: builds which are simple are good. There's nothing WRONG with a build that incorporates eight different prestige classes from seven different sources, and then tosses in feats from five more...but that build is going to be useful only to the people who have those sources, whereas the Druid 20 build that doesn't go outside of Core is useful to everybody. Sometimes, simplicity is worth more than raw power.
8. Tricking the DM is Bad.
We see a lot of "Help me trick my DM!" or "Help me make my DM cry!" requests on these boards. We see builds that are designed to look innocuous while at the same time being devastating to campaign balance. The idea is to lull the DM into allowing the character, then unleash its full power. Bad idea. Bad, BAD idea. At all times, two things should be borne in mind about the DM. One: he's in charge. If you try to trick him, he's totally within his rights to toss your character or YOU out of the game. Two: he's your friend. Trying to deceive your friends is bad. Be honest with your DM about what you want to do. If he says "No," deal with it. That's part of a DM's job. If you don't think he's going to say "Yes" to something, then trying to sneak it into the game on the sly is a sure way to make him mad.
9. Respect the parameters of the request.
This used to be a given, but people have been backsliding a lot lately. Someone comes on and says, "Hey, I'd like to play a Bard 4/Cleric 4. Can anyone help me optimize this? He immediately gets responses which boil down to, "Only an idiot would play that! You should be playing Pun-Pun, he's MUCH more powerful!" Sometimes they're more nicely phrased than this, other times they're not. The point is: people aren't offering him suggestions on how to make his character of choice better. They're telling him that he's "wrong" for playing that character, and that he should be playing a different character. The same goes for threads in which the poster explains the DM's house rules and restrictions at the beginning of the thread. More often than not, if these restrictions amount to more than "No infinite power at first level," someone will respond with the oh-so-helpful suggestion "Your DM sucks. Quit his game and never talk to him again."
I only wish that were hyperbole. It's word-for-word from a thread a while back. Optimization is about working within the rules to greatest effect. ANYONE can optimize in an environment with no restrictions. It takes skill to optimize where options are limited. Threads like these should be seen as an opportunity to demonstrate that skill...not belittle the poster or the DM.
10. If something seems too good to be true, it probably is.
I remember bounding onto the boards many moons ago, shortly after the first release of the Persistent Spell feat, to declare that I had discovered (ta da!) the UNBEATABLE COMBO. Since Time Stop was a Personal effect spell, it could be Persisted! (Oooh, aaah!) I couldn't imagine why nobody had thought of this before. Of course, as it turned out, LOTS of people had thought of this before. Within about five minutes, I was directed to a ruling that said, "You can't do it." I was disappointed, sure...but I accepted it and moved on. There are a LOT of folks here with a lot of knowledge of the rules. Some of 'em are a little scary. They love nothing better than to go over a new rulebook with a fine-toothed comb looking for hidden gems. Sometimes, a genuinely overlooked concept will turn up. The recent builds using Sanctum Spell are a good example. The feat's been around for a while, but nobody really looked at what could be done with it.
More often, though, if a seeming "rules loophole" is being ignored by the boards, it's because it's been hashed out in the past and found not to work. Perhaps there's something elsewhere in the rules that nullifies it; perhaps there was a clarification. Very occasionally, there's simply a board-wide agreement that the rule is wrong...as with the recent FAQ claiming that Polymorph allowed the use of templated forms. If it turns out that your discovery falls into this category, the best thing to do is accept it and move on. Maybe the next one won't.
So: there they are. Make of them what you will.