On 03/11/10 07:44, Jeremias Maerki wrote: > On 01.11.2010 21:29:14 Vincent Hennebert wrote: <snip/> >> • unimplemented methods should really throw an >> UnsupportedOperationException or be removed altogether. Their >> behaviour is likely to change once they are properly implemented, >> introducing regressions. Putting a comment in the Javadoc is not >> enough, especially since those methods can silently be called by >> intendedToRGB > > I guess it makes sense to make the methods private as long as they are > not used. Anyway, throwing UnsupportedOperationException doesn't work in > this case. Batik or FOP would crash whenever a color profile is > encountered that has a rendering intent that is not implemented. The > implemented way ensures that the user gets at least an approximate > result.
Whether an approximate result is preferable to an exception or not is a ‘business’ decision that I can’t make because I don’t know the underlying field well. What is the user likely to prefer? How acceptable is the fallback? If throwing an exception is impractical, is a warning desirable? Or is it likely to produce too much noise? Making the methods private seems to make sense indeed. <snip/> Vincent --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
