king's indian defense is a good opening and helps black acheive good king safety but white gains a superior center

I would prefer to play 4..d6 before castling. Why worry about 5.e5 if you don't have to?

 How do you know this is true? I was led to believe it was still a good defence and also understand it forms the basis of the Kings Indian Attack.

 If you don't play 4...d6 and e5 is played doesn't the dislodging of blacks knight wreck the defence?

 Reb thanks for you input into this discussion.

I don't see how Ne8 favours black. It seems like a retrograd move to me and hinders blacks development. Could you explain further?

let's quote Fischer from my 60 Memorable games

Letelier-Fischer

5 e5 "Weak. Letelier snapped at the chance to "take me out of the book" but this premature advance leaves white with all the responsibility of holding his overextended center pawns.

 That's great advice Reb. I'll try chessgames.com and post here if I find any other sources. Thank you.

5. ... Ne8 is a similar idea to dealing with the e5 variation of the Austrian Attack when playing the Pirc. Black tries to lure white's pawn centre forward to its death. John Nunn, who plays black in the game Tekoa posted, was (/is ?) a big exponent of the Pirc.

I guess 4...0-0 isn't played much because, as someone said, why worry about e5 if you don't have to, even if it's not good . Then again, it may be a good thing to give white this tempting option, as this tempts him to play an inferior move! But other than that, there are some lines where black may want to delay castling (more for psychological reasons, like fear of king safety) a little, like when white plays 5 Nge2 and tries an h4-h5 attack. I think castling kingside is technically fine in that line, but black has to be a bit careful, and he may prefer delaying castling just to avoid any sharp attack attempts by white, even if objectively they aren't so good.