I've seen this kind of code a lot recently:

try
{
int i = 0;
int x = 0;
Console.WriteLine(i / x);
}
catch (Exception ex)
{
Console.WriteLine(ex.Message);
}

This results in the following output:

Attempted to divide by zero.

Unless there's some compelling reason you need an ultra-terse version of the error, it's almost always better to use the provided Exception.ToString() method. Compare the difference:

System.DivideByZeroException: Attempted to divide by zero.
at ConsoleApplication1.Program.Main(String[] args)
in C:Program.cs:line 15

This also brings up an interesting corollary: any object you build should have a meaningful .ToString() method. I expect a proper string representation of what you are, not a meaningless echo of your class name!

One object that violates this horribly is DataSet. Let's say we have a DataSet containing a single DataTable. When you type DataSet.ToString(), what do you think should happen? Wait, wait, don't tell me. I'll tell you. This happens:

System.Data.DataSet

Useless. How about something that actually shows the object in string form?

+----------------------------------------------------------------+
| DataSet1                                                       |
+----------------------------------------------------------------+
| Table1                                                         |
+---------+-----------+------------------+------------+----------+
| field01 | field02   | field03          | field04    | field05  |
+---------+-----------+------------------+------------+----------+
|       1 | first     | NULL             | NULL       | NULL     |
|       2 | second    | another          | 1999-10-23 | 10:30:00 |
|       3 | a third   | more foo for you | 1999-10-24 | 10:30:01 |
+---------+-----------+------------------+------------+----------+

Seems perfectly logical to me, but you'll have to write your own custom ToString() implementation to get the behavior that should have been there in the first place.