The Castling Rights specify whether both sides are principally able to castle king- or queen side, now or later during the game - whether the involved pieces have already moved or in case of the rooks, were captured. Castling rights do not specify, whether castling is actually possible, but are a pre-condition for both wing castlings. Two bits per side are appropriate to store the castling rights, often one uses one nibble to store the rights for both sides inside a position object, a kind a array of four booleans.
In make move one has to consider that king-moves including castling itself reset both castling bits per side. Rook-moves from their original square, or captures of rooks on their original squares reset the appropriate castling bits per wing and side. Changed castling rights should be considered in the zobrist key of the position, to reflect the irreversibility of the otherwise reversible move, concerning repetitions. On the other hand, changed castling rights don't necessarily reset the halfmove clock regarding the fifty-move rule .
- Saving castling states and en passant history by Eric Lang, CCC, November 27, 2008 » En passant
- 0x88 FRC castle questions by Daniel Uranga, Winboard Forum, December 12, 2009 » Chess960
- Updating castling rights by Jan Brouwer, CCC, March 19, 2010
- ep and castle rights hashing by Natale Galioto, CCC, September 15, 2013 » En passant, Repetitions, Transposition Table
- 3rd repetition, a case where not cause castle rights... but by Luis Babboni, CCC, April 14, 2016 » Repetitions
- Enpass + Castling for Zorbist hashes by Andrew Grant, CCC, January 06, 2017 » En passant, Zobrist Hashing
- Open chess diary 41-60 52. 20 February: Castling rights in photographs and on servers by Tim Krabbé
- Computerschach - Eine Wette, die ich gerne verloren habe by Horst Wandersleben (German) 
- Mediocre Chess: [Bug Another little bug in castling rights] by Jonatan Pettersson, December 30, 2006
- Computerschach - Eine Wette, die ich gerne verloren habe by Horst Wandersleben (German)
- Dieter Bürßner found a game finished in a fifty-move rule draw, where castling occurred during the last fifty moves