--=============================================================================
--     Create and populate a table of sample data.
--     THIS IS NOT PART OF THE SOLUTION
--=============================================================================
 CREATE TABLE #yourtable
        (
        ClockIn  DATETIME,
        ClockOut DATETIME,
        UserID   INT
        )
 INSERT INTO #yourtable
        (ClockIn, ClockOut, UserID)
 SELECT '2008-02-21 09:54:25.000','2008-02-21 14:25:00.000',36 UNION ALL
 SELECT '2008-02-21 14:48:05.000','2008-02-21 17:00:00.000',36 UNION ALL
 SELECT '2008-02-22 12:55:19.000','2008-02-22 18:00:00.000',36 UNION ALL
 SELECT '2008-02-25 10:05:13.000','2008-02-25 15:14:50.000',36 UNION ALL
 SELECT '2008-02-25 15:44:52.000','2008-02-25 18:05:07.000',36 UNION ALL
 SELECT '2008-02-25 18:15:08.000','2008-02-25 18:20:31.000',36 UNION ALL
 SELECT '2008-02-25 18:30:33.000','2008-02-25 19:00:00.000',36 UNION ALL
 SELECT '2008-02-26 11:33:59.000','2008-02-26 17:50:00.000',36

--=============================================================================
--      Since SQL Server 2000 doesn't have a ROW_NUMBER() function, we have to
--      move the data to a temp table that has an IDENTITY column on it to make
--      the necessary row number for an ordered set.  Since the IDENTITY 
--      column will be used to reflect the order of data, we must create the
--      temp table first and then populate it instead of using SELECT/INTO.
--      This is pretty fast and certainly faster than doing a bunch of MIN
--      comparisons using correlated subqueries.
--=============================================================================
--===== This creates an empty temp table using SELECT/INTO... (cheater)
 SELECT IDENTITY (INT,1,1) AS RowNum,
        ClockIn,
        ClockOut,
        UserID
   INTO #MyWork
   FROM #yourtable
  WHERE 1 = 0 --This forces zero rows to be transferred.

--===== Use the same query with an order by to actually move the ordered data
 INSERT INTO #MyWork
        (ClockIn, ClockOut, UserID)
 SELECT ClockIn,
        ClockOut,
        UserID
   FROM #yourtable
  ORDER BY UserID, ClockIn

--===== Calculate and display break times that are <= 1 hour
 SELECT t1.UserID,
        t2.ClockOut,
        t1.ClockIn,
        DATEDIFF(mi,0,t1.ClockIn-t2.ClockOut) AS BreakMinutes
   FROM #MyWork t1,
        #MyWork t2
  WHERE t1.RowNum=t2.RowNum+1
    AND t1.UserID=t2.UserID
    AND DATEDIFF(mi,0,t1.ClockIn-t2.ClockOut) <=60 --1 Hour
  ORDER BY t1.UserID,t1.ClockIn

DROP TABLE #yourtable,#MyWork

--=============================================================================
--     Create and populate a table of sample data.
--     THIS IS NOT PART OF THE SOLUTION
--=============================================================================
 CREATE TABLE #yourtable
        (
        ClockIn  DATETIME,
        ClockOut DATETIME,
        UserID   INT
        )
 INSERT INTO #yourtable
        (ClockIn, ClockOut, UserID)
 SELECT '2008-02-21 09:54:25.000','2008-02-21 14:25:00.000',36 UNION ALL
 SELECT '2008-02-21 14:48:05.000','2008-02-21 17:00:00.000',36 UNION ALL
 SELECT '2008-02-22 12:55:19.000','2008-02-22 18:00:00.000',36 UNION ALL
 SELECT '2008-02-25 10:05:13.000','2008-02-25 15:14:50.000',36 UNION ALL
 SELECT '2008-02-25 15:44:52.000','2008-02-25 18:05:07.000',36 UNION ALL
 SELECT '2008-02-25 18:15:08.000','2008-02-25 18:20:31.000',36 UNION ALL
 SELECT '2008-02-25 18:30:33.000','2008-02-25 19:00:00.000',36 UNION ALL
 SELECT '2008-02-26 11:33:59.000','2008-02-26 17:50:00.000',36

--=============================================================================
--      Since SQL Server 2005 does have a ROW_NUMBER() function, don't have to
--      move data into a temp table to get the equivalent of an IDENTITY column.
--      And, since a CTE can be used more than once in the same query, we only
--      need 1 CTE.  All of it makes for some very short, easy to read, high
--      speed code.     
--=============================================================================
;WITH cteMyWork AS
(
 SELECT ROW_NUMBER() OVER (ORDER BY UserID,ClockIn) AS RowNum,
        ClockIn,
        ClockOut,
        UserID
   FROM #yourtable
)
 SELECT t1.UserID,
        t2.ClockOut,
        t1.ClockIn,
        DATEDIFF(mi,0,t1.ClockIn-t2.ClockOut) AS BreakMinutes
   FROM cteMyWork t1,
        cteMyWork t2
  WHERE t1.RowNum=t2.RowNum+1
    AND t1.UserID=t2.UserID
    AND DATEDIFF(mi,0,t1.ClockIn-t2.ClockOut) <=60 --1 Hour 

DROP TABLE #yourtable

DECLARE	@Sample TABLE
	(
		ClockIn DATETIME,
		ClockOut DATETIME,
		UserID INT
	)

INSERT	@Sample
        (
		ClockIn,
		ClockOut,
		UserID
	)
SELECT	'2008-02-21 09:54:25', '2008-02-21 14:25:00', 36 UNION ALL
SELECT	'2008-02-21 14:48:05', '2008-02-21 17:00:00', 36 UNION ALL
SELECT	'2008-02-22 12:55:19', '2008-02-22 18:00:00', 36 UNION ALL
SELECT	'2008-02-25 10:05:13', '2008-02-25 15:14:50', 36 UNION ALL
SELECT	'2008-02-25 15:44:52', '2008-02-25 18:05:07', 36 UNION ALL
SELECT	'2008-02-25 18:15:08', '2008-02-25 18:20:31', 36 UNION ALL
SELECT	'2008-02-25 18:30:33', '2008-02-25 19:00:00', 36 UNION ALL
SELECT	'2008-02-26 11:33:59', '2008-02-26 17:50:00', 36

SELECT		d.UserID,
		DATEADD(DAY, DATEDIFF(DAY, '19000101', d.minTime), '19000101') AS theDay,
		d.theWork / 3600.0 AS WorkHours,
		(DATEDIFF(SECOND, d.minTime, d.maxTime) - d.theWork) / 60.0 AS BreakMinutes
FROM		(
			SELECT		UserID,
					MIN(ClockIn) AS minTime,
					MAX(ClockOut) AS maxTime,
					SUM(DATEDIFF(SECOND, ClockIn, ClockOut)) AS theWork
			FROM		@Sample
			GROUP BY	UserID,
					DATEDIFF(DAY, '19000101', ClockIn)
		) AS d

SELECT		d.UserID,
		DATEADD(DAY, DATEDIFF(DAY, '19000101', d.minTime), '19000101') AS theDay,
		CONVERT(CHAR(8), DATEADD(SECOND, d.theWork, '19000101'), 108) AS Work,
		CONVERT(CHAR(8), DATEADD(SECOND, DATEDIFF(SECOND, d.minTime, d.maxTime) - d.theWork, '19000101'), 108) AS [Break]
FROM		(
			SELECT		UserID,
					MIN(ClockIn) AS minTime,
					MAX(ClockOut) AS maxTime,
					SUM(DATEDIFF(SECOND, ClockIn, ClockOut)) AS theWork
			FROM		@Sample
			GROUP BY	UserID,
					DATEDIFF(DAY, '19000101', ClockIn)
		) AS d