The DISTINCT keyword is often a question of debate.  Many ask “how to do a distinct only on the first few columns of a select statement ?”.  There is a fundamental flaw in this question, since DISTINCT is always applied to all columns in the SELECT.  Let’s look at an example:

create table #JobHistory (
  Title nvarchar(50),
  FirstName nvarchar(50),
  LastName nvarchar(50),
  StartDate datetime,
  EndDate datetime
)

Insert into #JobHistory (Title, FirstName, LastName, StartDate, EndDate)
Values
('Developer', 'Brian', 'Carter', '2002-02-02', '2004-04-04'),
('DBA', 'Brian', 'Carter', '2004-04-04', '2008-08-08'),
('Developer', 'Brian', 'Carter', '2008-08-08', '2010-10-10')

SELECT Title, FirstName, LastName, 
  CONVERT(VARCHAR(10), StartDate, 111) AS StartDate, 
  CONVERT(VARCHAR(10), EndDate, 111) AS EndDate
FROM #JobHistory

The result, from SQL Server:

Title       FirstName LastName StartDate  EndDate
Developer   Brian     Carter   2002/02/02 2004/04/04
DBA         Brian     Carter   2004/04/04 2008/08/08
Developer   Brian     Carter   2008/08/08 2010/10/10

Often, what is needed is a list of distinct titles for a given person.
Let’s say we need a list:

Title     FirstName LastName
Developer Brian     Carter
DBA       Brian     Carter

The query would look like:

SELECT DISTINCT Title, FirstName, LastName
FROM #JobHistory

As you can see, the distinct was applied to all columns in the select.
The same result can be achieved by using GROUP BY as well:

SELECT Title, FirstName, LastName
FROM #JobHistory
GROUP BY Title, FirstName, LastName

The issue I have with Group By, is all columns in the Select must be listed (or use aggregate function like MAX, MIN, SUM, …).  This subject is very interesting.  I found another good write up on Jeffs blog.