Break the Rules
I consider myself an upper intermediate player. My rating fluctuates between 1700-2050, and I've gone through GM games usually in the English and the Dragon where they move the bishop in front of the pawn usually on d3 in the English and e6 in the dragon. Also I've seen GM's move there knights on the rim in their opening moves. This is inconsistent with what I was taught. My question is when is this good, and when is this bad?
When you mentioned the bishop in front of the pawn on d3, I am going to assume you actually meant e3 since that is more common in the English opening. I personally have taught several beginners never to develop a piece and block one of their central pawns. Before I answer your question let me begin with the reasons behind some of these basic strategies.
Why should you not develop one of your pieces in front of a central pawn and block it from advancing?
One of the basic concepts that we are taught in chess is the idea of space. How do we gain space in chess? We certainly do not achieve space by moving our pieces forward, it is the pawn advances that give you space. Think of it like this, the pawn formation is like the fortress of your castle. You can call anything inside your pawn formation "your space," but the space outside can always be contested by your opponent. This does not mean you can randomly push your pawns forward and claim the territory. The advanced pawn should be backed up well to maintain the space that you intend to gain. Now, that you know you want to gain space, where would you want to do that? Obviously, it has to be in the most important part of the chess board, the center. After reading through some of the basics here, I think it should become evident for any chess player why one should not block a central pawn from advancing. The role of the 'e' and 'd' pawns in most games are either to advance themselves in pursuit of space or protect the other advanced pawn to maintain the space and control the center.
Given that this rule holds true for most chess openings, it is easier to teach someone to just follow the rule blindly to begin with. But, as we know every rule in life as well as chess has exceptions; after a point it is important to understand these exceptions as well.
This is a position from the Dragon variation of the Sicilian defense and as you had mentioned in your question, black just developed his bishop to e6 blocking his own central pawn.
Why is this an exception to the rule? The answer is in understanding the key to the position. In this case, the pawn on d6, even though not very advanced, is the one that is holding the space and controlling the center for black. Pushing the pawn to e5 blocks the bishop on g7 and also gives away an outpost for white on d5. Trying to advance e6 is no good either as it weakens his d6 pawn, which will be exposed to tremendous pressure along the semi open 'd' file. Also, in this case the black bishop will not be in e7 to protect the d6 pawn, which is the case in other variations of the Sicilian defense. The final question before you make your decision would be, “Can your opponent expand further if you do not?” In this case white trying to play e5 is not very practical. White has already committed his pawn to f3, so f4 and e5 ideas are virtually impossible. After taking all these strategic points into consideration, we can safely conclude that this position is indeed an exception to the rule and advancing the 'e' pawn in fact makes things worse for black. Hence, blocking the pawn is irrelevant in this situation and on top of that the black bishop is very well placed on e6 to participate in any future attack on the white king.
This position is from the English opening. White just developed his bishop to e3. Even though there were several other good options for white, this particular development is considered to be fine as well. Now, let us check some strategic points and see why this one will fit into the category of exception.
Why didn't white think of advancing the 'e' pawn before placing his bishop on e3? First reason would be that it is blocking white's light squared bishop on g2. Secondly e4 is very important if black was trying to break free with d5, but in this case white has very good control of the 'd' file and black just cannot think about d5. Advancing the 'e' pawn to e3 does not do any good in this case, so it is much better to leave the pawn where it is and find good squares for the pieces.
If you notice one common point between the two positions, both sides that broke this rule had their bishop fianchettoed before they broke the rule. Think about it, if you had your king side bishop undeveloped sitting on f1 or f8, would it be good to block the 'e' pawn? That would just make things very difficult for your development.
The second part of your question is a little more straight-forward to understand. As a general principle, knights on the rim are grim. It is very easy to understand the logic behind this simple rule. Place a knight in the middle of the board and count the number of squares it can reach, it should be 8. Now remove the knight and try placing it on one of the edges, let us say a5. You will find that the knight can only move to 4 squares from there. Try placing it on a8 and it will make things even worse. This time the knight can only move to 2 squares. This simply tells us that the activity of a knight gets curtailed more and more as it gets closer to the edge. This is why in the beginning of a game every one is advised to develop their knights towards the center and not towards the edges. There is one main exception to this rule. You can place a knight on the rim if it is just en-route to a different square soon. Very rarely the knight would be just well placed on the rim itself.
Take a look at this example, a common position from the Benoni defense
Black just developed his knight to a6, something we are not taught to do in a game. Let us see what could have prompted black to put his knight in such an awkward square. The knight on b8 had two choices, to head to d7 or a6. Instead of looking at where the knight would go to in the next move, let us think about where would we want this knight to be and where would we like our light squared bishop to be eventually. Black's main plan in this position is to create a pawn break at b5. Middle game planning involves pawn breaks and pawn breaks mean that you need your pieces backing up the breaks. Keeping our plan in mind we know that a knight would be very handy on c7 and the bishop on d7 to back up the b5 pawn break. So just as I had mentioned earlier, the knight is not meant to remain on the rim, it is just passing through a bad square to get to a different square from which it can support your plan better.
I hope this answered your question regarding the basic rules and their exceptions. In my opinion breaking the rules is very important, because that is what differentiates a man from a machine. But, to be able to break a rule, you should know what the rule stands for in the first place. When you understand the nuances better, you will automatically understand when that particular rule is not needed and when it is absolutely necessary.