diff --git a/cwFlowTypes.h b/cwFlowTypes.h
index 221f0ad..7a01b46 100644
--- a/cwFlowTypes.h
+++ b/cwFlowTypes.h
@@ -468,7 +468,7 @@ namespace cw
     void           var_destroy( variable_t* var );
 
     // Channelizing creates a new var record with an explicit channel index to replace the
-    // automatically generated variable whose channel index is set to  'all'.
+    // automatically generated variable whose channel index is set to  'kAnyChIdx'.
     rc_t           var_channelize( proc_t* proc, const char* var_label, unsigned sfx_id, unsigned chIdx, const object_t* value_cfg, unsigned vid, variable_t*& varRef );
 
     // Wrapper around call to var->proc->members->value()
diff --git a/cwObject.h b/cwObject.h
index 8bc78d7..d58ee20 100644
--- a/cwObject.h
+++ b/cwObject.h
@@ -32,7 +32,9 @@ namespace cw
 
    kHexFl      = 0x10000000,
    kIdentFl    = 0x20000000,
-   kOptFl      = 0x40000000
+   
+   kOptFl      = 0x40000000,
+   kReqFl      = 0x00000000,
    
   };
 
@@ -304,10 +306,13 @@ namespace cw
 
     
     // readv("label0",flags0,v0,"label1",flags0,v1, ... )
-    // Use kOptFl for optional fields.
+    // Use kReqFl/kOptFl for required/optional fields.
     // Use kListTId and kDictTId to validate the type of container fields.
     // In general it should not be necessary to validate numeric and string types because
     // they are validated by virtue of being converted to the returned value.
+    // Note that readv() assumes that the list of possible fields given as input is complete
+    // and any fields that it finds which are not in the list are not valid. This
+    // validity check is the main difference between readv() and getv()/getv_opt().
     template< typename T0, typename T1, typename... ARGS >
     rc_t readv( T0 label, unsigned flags, T1& valRef, ARGS&&... args ) const
       { return _readv(nullptr, label,flags,valRef,args...); }