On 3 October 2012 18:24, Jörg Schaible <joerg.schai...@gmx.de> wrote: > Matt Benson wrote: > >> Urgh; I find these method names rather painful. Why wouldn't we >> simply provide endianness and bit ordering as enums, and parameterize >> accordingly? > > Because the algorithm is different (although similar) every time and not all > combinations are implemented? > > Honestly, we would have to use in most cases a switch/case statement > internally anyway and either throw UnsupportedOperationException for the > unimplemented cases or someone would have to implement it ... ;-)
I think the enum-based solution is much more elegant; the JavaDoc could then contain a table demonstrating which combinations are supported. The better of two evils, IMO. I'm always in favour of methods that can be easily understood from a cursory glance. It bodes well for easy code maintenance. Duncan --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
