diff --git a/lib/galaxy/jobs/__init__.py b/lib/galaxy/jobs/__init__.py
index f67980da775..02af13bee50 100644
--- a/lib/galaxy/jobs/__init__.py
+++ b/lib/galaxy/jobs/__init__.py
@@ -230,15 +230,8 @@ class JobWrapper( object ):
             data = dataset_assoc.dataset
             if data.state == data.states.FAKE:
                 data.deleted = True
+        # TODO
         # validate output datasets
-        for out_dataset in job.output_datasets:
-            # file must exist in order to validate it
-            if os.path.exists(out_dataset.dataset.file_name):
-                errors = out_dataset.dataset.datatype.validate(out_dataset.dataset)
-                for error in errors:
-                    out_dataset.dataset.add_validation_error(
-                        model.ValidationError( message=str( error ), err_type=error.__class__.__name__, attributes=str( error.__dict__ ) ) )
-                    print str( error )
                         
         mapping.context.current.flush()
         log.debug('job ended, id: %d' % self.job_id )
diff --git a/lib/galaxy/model/__init__.py b/lib/galaxy/model/__init__.py
index b4080af04b0..f83f538a45c 100644
--- a/lib/galaxy/model/__init__.py
+++ b/lib/galaxy/model/__init__.py
@@ -254,7 +254,10 @@ class Dataset( object ):
 
     def add_validation_error( self, validation_error ):
         self.validation_errors.append( validation_error )
-    
+
+    def extend_validation_errors( self, validation_errors ):
+        self.validation_errors.extend(validation_errors)
+
     # FIXME: sqlalchemy will replace this
     def _delete(self):
         """Remove the file that corresponds to this data"""
@@ -302,4 +305,4 @@ class GalaxySession( object ):
 class GalaxySessionToHistoryAssociation( object ):
     def __init__( self, galaxy_session, history ):
         self.galaxy_session = galaxy_session
-        self.history = history
\ No newline at end of file
+        self.history = history
diff --git a/lib/galaxy/tools/actions/upload.py b/lib/galaxy/tools/actions/upload.py
index 4b9a6e45f81..1ee2586c041 100644
--- a/lib/galaxy/tools/actions/upload.py
+++ b/lib/galaxy/tools/actions/upload.py
@@ -1,6 +1,7 @@
 import os, shutil, urllib, StringIO
 from galaxy import datatypes, jobs
 from galaxy.datatypes import sniff
+from galaxy import model
 
 class UploadToolAction( object ):
     """
@@ -82,6 +83,12 @@ class UploadToolAction( object ):
         if isinstance( data.datatype, datatypes.interval.Interval ):
             if data.missing_meta():
                 data.extension = 'tabular'
+
+        # validate incomming data
+        for error in data.datatype.validate( data ):
+            data.add_validation_error( 
+                model.ValidationError( message=str( error ), err_type=error.__class__.__name__, attributes=str( error.__dict__ ) ) )
+            
         trans.history.add_dataset( data )
         trans.app.model.flush()
         return data
diff --git a/scripts/validate.py b/scripts/validate.py
index 97f911d1030..dd8cdc48e6d 100644
--- a/scripts/validate.py
+++ b/scripts/validate.py
@@ -5,7 +5,7 @@ Validate a dataset based on extension a metadata passed in on the
 command line.  Outputs a binhex'd representation of the exceptions.
 
 usage: %prog input output
-    -c, --cols=N,N,N,N: column metadata, in the case of GFF, BED, or intervals
+    -m, --metadata=N: base64 pickeled metadata
     -x, --ext=N: extension as understood by galaxy
 """
 from cookbook import doc_optparse
@@ -25,12 +25,8 @@ def main():
     data = model.Dataset( extension=extension, id=int( args[0] ) )
     data.file_path = "/home/ian/trunk/database/files/"
     
-    if options.cols:
-        cols = options.cols.split(",")
-        data.metadata.chromCol = cols[0]
-        data.metadata.startCol = cols[1]
-        data.metadata.endCol = cols[2]
-        data.metadata.strandCol = cols[3]
+    if options.metadata:
+        data.metadata = util.string_to_object( options.metadata )
 
     errors = data.datatype.validate( data )
     print util.object_to_string(errors)