diff --git a/uf-bare b/uf-bare
index 0a9c40b..c4916a3 100755
--- a/uf-bare
+++ b/uf-bare
@@ -71,6 +71,6 @@ done
 # Do the work
 
 awk -b -O -v F="$HDRSFILE" '
-	NR%2==1 && F { print > "'"$HDRSFILE"'" }
+	NR%2==1 && F { print > F }
 	NR%2==0' "$@"
 
diff --git a/uf-dress b/uf-dress
index c737b96..6e85a5b 100755
--- a/uf-dress
+++ b/uf-dress
@@ -29,18 +29,22 @@ usage_exit() {
 	echo
 	echo "Usage: $(basename $0) [OPTIONS] [FILE] ..."
 	echo
-	echo "  Insert headers to turn bare sequence FILE(s) into valid unfasta.  If no FILE or"
-	echo "  FILE is '-' read from standard input.  Write unfasta to standard output."
+	echo "  Insert headers to turn bare sequences into valid unfasta.  Reads each line from"
+	echo "  each FILE and writes it to standard output preceded by a header line.  If no 
+	echo "  FILE is present or FILE is '-' read from standard input."
 	echo
 	echo "  Options"
-	echo "   -r HDRSFILE   Optional file to read header lines from.  If not specified then"
-	echo "                 default headers 'lcl|NUM' will be generated."
+	echo "   -r HDRSFILE  Optional file to read header lines from.  If not specified or if"
+	echo "                this file has fewer header lines than sequences, then dummy header"
+	echo "                lines are generated for the remaining sequences."
 	echo
 	echo "  The HDRSFILE could be a file previously written by 'uf-bare', or output from" 
 	echo "  the 'uf-hdrs' filter.  Tip: use bash process substitution to avoid temp files:"
 	echo
 	echo '   $ uf file.fna | uf-bare | ..processing.. | uf-dress -r <(uf file.fna | uf-hdrs)'
 	echo
+	echo "  See also: 'uf-map' which implements this idiom in a single command."
+	echo
 	exit ${1:-1}
 }
 
@@ -70,7 +74,9 @@ done
 # Do the work
 
 awk -b -O -v F="$HDRSFILE" '{
-	if (F) getline HDR < F; else HDR = ">lcl|" NR
+	HDR = ""
+	if (F) getline HDR < F
+	if (!HDR) HDR = ">lcl|" NR " Dummy header " NR
 	print HDR
 	print
 	} 
diff --git a/uf-map b/uf-map
index f25d17a..f51ece2 100755
--- a/uf-map
+++ b/uf-map
@@ -62,7 +62,39 @@ done
 OPERATION="$1"
 shift
 
-# AWK it
+# Do it 
 
-awk -b -O "NR%2==1 { print }; NR%2==0 { print | \"$OPERATION\" }"
+# FAIL 1: using pipe within awk
+# Fails because of race: external operation is buffering its output, so lines do not come out in the right order.
+#awk -b -O -v OPER="sh -c '$OPERATION'" '
+#	NR%2==1 { print }
+#	NR%2==0 { print | OPER }
+#	{ fflush() }' "$@"
+
+# FAIL 2: using a FIFO between uf-bare and uf-dress
+# Fails because of deadlock: uf-bare blocks on FIFO until uf-dress consumes,
+# but uf-dress is waiting for standard input before consuming from FIFO.
+#FIFO="/tmp/$(basename "$0").$$"
+#mkfifo "$FIFO" 
+#./uf-bare -f -w "$FIFO" | ./uf-dress -r "$FIFO"
+#RETVAL=$?
+#rm -f $FIFO
+
+# Maybe using paste to pull things together?
+
+# Set pipefail to have the errorcode of the rightmost failing filter (OPERATION)
+[ -z "$BASH" ] set +o pipefail
+
+# Make a FIFO to hold the headers and from which paste merges them with OPERATION standard output
+FIFO="/tmp/$(basename "$0").$$"
+mkfifo "$FIFO" 
+
+# Run the pipeline
+awk -b -O -v FIFO="$FIFO" 'NR%2==1 { print > FIFO }; NR%2==0' "$@" | $OPERATION | paste -d '\n' "$FIFO" -
+RETVAL=$?
+
+# Clean up - @TODO@ set a shell trap, is cleaner
+rm -f $FIFO
+
+exit $RETVAL