diff --git a/VERSION b/VERSION
index 258fb691..34c73233 100644
--- a/VERSION
+++ b/VERSION
@@ -1 +1 @@
-Version 1.99.4 (2011-12-16 09:44:46) stable
+Version 1.99.4 (2011-12-16 09:55:58) stable
diff --git a/gluon/contrib/timecollect.py b/gluon/contrib/timecollect.py
new file mode 100644
index 00000000..69bd8c0c
--- /dev/null
+++ b/gluon/contrib/timecollect.py
@@ -0,0 +1,95 @@
+# Only Python 2.6 and up, because of NamedTuple.
+import time
+from collections import namedtuple
+Score = namedtuple('Score', ['tag', 'stamp'])
+
+class TimeCollector(object):
+    def __init__(self):
+        '''The first time stamp is created here'''
+        self.scores = [Score(tag='start',stamp=time.clock())]
+    def addStamp(self, description):
+        '''Adds a new time stamp, with a description.'''
+        self.scores.append(Score(tag=description, stamp=time.clock()))
+    def _stampDelta(self, index1, index2):
+        '''Private utility function to clean up this common calculation.'''
+        return self.scores[index1].stamp - self.scores[index2].stamp
+    def getReportItems(self, orderByCost=True):
+        '''Returns a list of dicts. Each dict has 
+            start (ms), 
+            end (ms), 
+            delta (ms), 
+            perc (%),
+            tag (str)
+        '''
+        self.scores.append(Score(tag='finish', stamp=time.clock()))
+        total_time = self._stampDelta(-1, 0)
+        data = []
+        for i in range(1, len(self.scores)):
+            delta = self._stampDelta(i, i-1)
+            if abs(total_time) < 1e-6:
+                perc = 0
+            else:
+                perc = delta / total_time * 100
+            data.append(
+                dict(
+                    start = self._stampDelta(i-1, 0) * 1000,
+                    end = self._stampDelta(i, 0) * 1000,
+                    delta = delta * 1000,
+                    perc = perc,
+                    tag = self.scores[i].tag
+                    ) 
+                )
+        if orderByCost:    
+            data.sort(key=lambda x: x['perc'], reverse=True)
+        return data
+    def getReportLines(self, orderByCost=True):
+        '''Produces a report of logged time-stamps as a list of strings.
+        if orderByCost is False, then the order of the stamps is
+        chronological.'''
+        data = self.getReportItems(orderByCost)
+        headerTemplate = '%10s | %10s | %10s | %11s | %-30s'
+        headerData = ('Start(ms)', 'End(ms)', 'Delta(ms)', 'Time Cost',
+                      'Description')        
+        bodyTemplate = '%(start)10.0f | %(end)10.0f | %(delta)10.0f |' \
+          + ' %(perc)10.0f%% | %(tag)-30s' 
+        return [headerTemplate % headerData] + [bodyTemplate % d for d in data]
+    def getReportText(self, **kwargs):
+        return '\n'.join(self.getReportLines(**kwargs))
+    def restart(self):
+        self.scores = [Score(tag='start',stamp=time.clock())]
+            
+if __name__=='__main__':
+    print('')
+    print('Testing:')
+    print('')
+    
+    # First create the collector
+    t = TimeCollector()
+    x = [i for i in range(1000)]
+    # Every time some work gets done, add a stamp
+    t.addStamp('Initialization Section')
+    x = [i for i in range(10000)]
+    t.addStamp('A big loop')
+    x = [i for i in range(100000)]
+    t.addStamp('calling builder function')
+    # Finally, obtain the results
+    print('')
+    print(t.getReportText())
+    # If you want to measure something else in the same scope, you can
+    # restart the collector.
+    t.restart()
+    x = [i for i in range(1000000)]
+    t.addStamp('Part 2')
+    x = [i for i in range(1000000)]
+    t.addStamp('Cleanup')
+    # And once again report results
+    print('')
+    print(t.getReportText())
+    t.restart()                         
+    for y in range(1, 200, 20):
+        x = [i for i in range(10000)*y]
+        t.addStamp('Iteration when y = ' + str(y))
+
+    print('')
+    # You can turn off ordering of results
+    print(t.getReportText(orderByCost=False))