DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT <http://nagoya.apache.org/bugzilla/show_bug.cgi?id=20774>. ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND INSERTED IN THE BUG DATABASE.
http://nagoya.apache.org/bugzilla/show_bug.cgi?id=20774 [math] gamma test cases ------- Additional Comments From [EMAIL PROTECTED] 2003-06-17 02:37 ------- > Even though I somewhat disagree with this implementation What don't you like? > If we want to start throwing IllegalArgumentException for these situations All the IAE have been removed. I thought this is want you desired. > I'd be opposed to writing Math.sqrt() and throwing an illegal > argument for the parameter of -1.0, but in Gamma it might make > some sense. Why? The routines in Gamma are no more complex than anything in java.lang.Math. I thought you wanted simple methods to return NaN, like the methods in java.lang.Math, instead of throwing IAE. > Throwing too many runtime exception really just makes a library > somewhat unusable. Then I've made the Gamma methods more usable by removing some of the runtime exceptions. The only ones that remain are the convergence exceptions, which I believe you said were appropriate. IMO, the guidelines governing NaN and runtime exceptions are too subjective which results in things like the above: contributors submitting things they believe conform to the guidelines but are in fact in violation of the guidelines according to the powers that be. It might be better to explicitly state the guidelines on when and where NaN's should be used by documenting them and adding them the the developer portion of the site. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]