A Bug or not a bug?

*To*: mathgroup at smc.vnet.net*Subject*: [mg74883] A Bug or not a bug?*From*: "dimitris" <dimmechan at yahoo.com>*Date*: Tue, 10 Apr 2007 05:12:45 -0400 (EDT)

This appears in another forum. (Actually the third integral below. The discoverer regards it as a bug. The first two integrals are part of my response to him). Integrate[Exp[x], {x, -Infinity, 0}] (*correct*) 1 Integrate[Exp[x], {x, 0, Infinity}] (*correct*) Integrate::idiv: Integral of E^x does not converge on {0, Infinity}. Integrate[Exp[x], {x, 0, Infinity}] Integrate[Exp[x], {x, -Infinity, 0}, Assumptions -> x < 0] (*incorrect*) Integrate::idiv: Integral of E^(-x) does not converge on {0,Infinity}. Integrate[E^x, {x, -Infinity, 0}, Assumptions -> x < 0] Since my experience about what must be regarded as a bug/feature is not so big as others in this forum I would like to see what you think about the last performance of Mathematica. I personally don't regard it as a serious bug. It is like trying to "cheat" Mathematica. I am sure there are a great many ways to make built in functions to give wrong results! And what is more important the unecessary setting of Assumptions caused it! Your Opinion? Dimitris