by Kurt » Wed Apr 25, 2007 9:59 pm
You can really even apply the same logic to avoid integration by parts in some simple cases (although arguably most of those "simple" cases can employ tabular integration to "avoid" it as well).
Good example: [unparseable or potentially dangerous latex formula].
You could use integration by parts, but think about it: what type of function could possibly contain a [unparseable or potentially dangerous latex formula] in the derivative? Obviously something that allows for the product rule, in which the [unparseable or potentially dangerous latex formula] must remain. But this other function needs to drop to 1 (so that just [unparseable or potentially dangerous latex formula] remains) when you do that part of the product rule, so the other function multiplied on needs to be [unparseable or potentially dangerous latex formula]. So you have [unparseable or potentially dangerous latex formula]...but the derivative of that is [unparseable or potentially dangerous latex formula]...so you also need something that, when you take the derivative, cancels out the 1 out front - which is [unparseable or potentially dangerous latex formula].
So [unparseable or potentially dangerous latex formula].
(...It ends up being a lot quicker to do intuitively then to explain doing it intuitively. :P Either way, it ends up being handy on all sorts of problems when you don't want to bother writing down stuff - stuff like this can easily be reasoned out mentally.)