Look at the comments you'll see it.
castling on a file

according to polish Wikipedia http://pl.wikipedia.org/wiki/Roszada_Pama-Krabb%C3%A9go
Tim Krabbe publicised the idea in 1972, and FIDE made it illegal the same year
The whole concept of castling on a file is dumb.
You may only castle if neither of the king nor castle has moved. In order to get to the other side of the board the pawn made 5 or 6 moves. When you promote a piece in chess it is still the SAME piece but it just moved up in rank.
Not to mention the fact that it would have no real use in a game. If your king is safe enough to be left on an open central file w/o ever moving long enough for you to march a pawn all the way down to the other side of the board... Well maybe it's time you find a tougher opponent.

No, it's not dumb;
1. When we promote a pawn on a real chess board, we remove the pawn first, and then place a rook on its place. So it isn't the same piece.
2. It actually was proposed as a solution to a problem in 1972 which is a two move mate. http://www.futilitycloset.com/2009/12/11/outside-the-box/
3. But to stop this fantastic castling move, FIDE appended the word 'rank' to the rule, which excludes so called 'vertical castling'. I couldn't find if vertical castling is legit today, or not.
In the FIDE rule book it is mentioned that for castling, rook and king have to be on the same rank....refer 3.8 http://www.fide.com/component/handbook/?id=124&view=article .However, i have seen a game somewhere, b/w two GMs where castling was done along a file. The situation i remember is as follows:
white king is one e1,previously unmoved.while pawn is on e7 abt to be promoted. black king is on h2.the pawn on e7 was promoted to rook after white's move to e8.And now, since the king and rook are previously unmoved he castled king to e3 while the rook reached e2 imposing a threat to king.
I want to know if the move is legitimate,because such a thing is not mentioned in the rule book.