diff --git a/src/cvtDctsdr/Makefile.Unix b/src/cvtDctsdr/Makefile.Unix
index 18ebe0221..3508124bb 100644
--- a/src/cvtDctsdr/Makefile.Unix
+++ b/src/cvtDctsdr/Makefile.Unix
@@ -5,12 +5,14 @@ include $(EPICS)/config/CONFIG_BASE
 USR_LDLIBS =  -lCom
 USR_LDFLAGS = -L.
 
+
 DEPLIBS_BASE = $(EPICS_BASE_LIB)
 
 DEPLIBS = \
 	$(DEPLIBS_BASE)/libCom.a
 
 SRCS.c = \
+	sf2dbYacc.c\
 	../sdr2gblmenu.c\
 	../sdr2driver.c\
 	../sdr2device.c\
@@ -24,7 +26,9 @@ OBJS = \
 	sdr2recordtype.o \
 	dbStaticLib.o
 
-TARGETS = sdr2gblmenu sdr2driver sdr2device sdr2recordtype
+MAN1 = sf2db.1
+
+PROD = sf2db sdr2gblmenu sdr2driver sdr2device sdr2recordtype 
 
 include $(EPICS)/config/RULES.Unix
 
@@ -39,3 +43,12 @@ sdr2device: sdr2device.o dbStaticLib.o
 
 sdr2recordtype: sdr2recordtype.o dbStaticLib.o
 	$(LINK.c) -o $@  sdr2recordtype.o dbStaticLib.o  $(LDLIBS)
+
+sf2db: sf2dbYacc.o
+	$(LINK.c) -o $@ sf2dbYacc.o
+
+sf2dbYacc.o: sf2dbLex.c
+
+clean::
+	/bin/rm -f sf2dbYacc.c sf2dbLex.c
+
diff --git a/src/cvtDctsdr/sf2db.1 b/src/cvtDctsdr/sf2db.1
new file mode 100644
index 000000000..7271da1ed
--- /dev/null
+++ b/src/cvtDctsdr/sf2db.1
@@ -0,0 +1,36 @@
+.\" @(#)
+.TH SF2DB 1 "04 Nov 1993"
+.SH NAME
+sf2db \- make a new ascii database ".db" file from an old dct short form report
+.SH SYNOPSIS
+.B sf2db
+.B name
+<
+.B short_form_rpt
+>
+.B db_file
+.SH AVAILABILITY
+The tool in available under Unix from the EPICS add_on directory.
+.SH DESCRIPTION
+.LP
+This tool reads a dct short form report file from standard in and
+writes a ".db" (new ascii database format) file to standard out.  A
+.B name
+must be given for the database.  The name specified is used to name
+the ".database" file when the ".db" file is converted to a binary ".database".
+.sp
+The output of this tool can be imported to gdct(1) for editting.
+.SH "EXAMPLE USAGE"
+.LP
+sf2db test < old.short.form > new.db
+.sp
+This command will read old.short.form and produce the file new.db.  When
+new.db is fed into db2database(1), a binary database file will be created
+named test.database.
+.SH "SEE ALSO"
+.BR gdct(1),
+.BR db2database(1),
+.BR dbLoadRecords(3),
+.BR dbfile(5),
+.sp
+.B "GDCT User's Manual"
diff --git a/src/cvtDctsdr/sf2dbLex.l b/src/cvtDctsdr/sf2dbLex.l
new file mode 100755
index 000000000..d13bd0a8b
--- /dev/null
+++ b/src/cvtDctsdr/sf2dbLex.l
@@ -0,0 +1,26 @@
+b	[a-zA-Z0-9_]
+a	[ \t]
+d	[a-zA-Z0-9_\,\./\*#\{\}\[\]%:;!|\'\-&\(\)@\?\+<>=$\^\~]
+
+%{
+%}
+
+%%
+
+^[A-Z0-9_]+	{ yylval.Str=(char *)malloc(strlen(yytext)+1);
+			strcpy(yylval.Str,yytext);
+			return(FIELD); }
+
+^"PV:"		{ return(PV); }
+"Type:" 	{ return(TYPE); }
+""		{ return(CLOSE); }
+"$$end"		{ return(CLOSE); }
+
+{d}+		{ yylval.Str=(char *)malloc(strlen(yytext)+1);
+			strcpy(yylval.Str,yytext);
+			return(WORD);   }
+
+.			;
+\n			{ line_num++; }
+
+%%
diff --git a/src/cvtDctsdr/sf2dbYacc.y b/src/cvtDctsdr/sf2dbYacc.y
new file mode 100755
index 000000000..1739c66d8
--- /dev/null
+++ b/src/cvtDctsdr/sf2dbYacc.y
@@ -0,0 +1,151 @@
+%{
+#include <stdio.h>
+#include <memory.h>
+#include <string.h>
+#include <epicsVersion.h>
+
+#undef PUKE_FACE
+#if EPICS_VERSION<4
+#if EPICS_REVISION<13
+#if EPICS_MODIFICATION<1 && EPICS_UPDATE_LEVEL<12
+#define PUKE_FACE
+#endif
+#endif
+#endif
+
+/* kludge for buggy sun lex/yacc.  exploits the fact that we know the union */
+/* below will be given the name YYSTYPE.  done so that ifndef YYSTYPE  */
+/* (as it appears in pdb.c) fails */
+/* #define YYSTYPE OUR_YYSTYPE */
+
+int line_num;
+char Field_val[1000];
+
+%}
+
+%start sf2db
+
+%token <Str> WORD
+%token <Str> FIELD
+%token <Str> TYPE
+%token <Str> PV CLOSE
+
+%union
+{
+    int	Int;
+    char Char;
+    char *Str;
+    double Real;
+}
+
+%%
+
+sf2db:	crap_head records closer
+	{ printf("\n"); }
+	;
+
+closer:	
+	| CLOSE
+	;
+
+crap_head: 
+	| crap_header CLOSE
+	| CLOSE
+	| crap_header
+	;
+
+crap_header: crap_header WORD
+	| WORD
+	;
+
+records: records record
+	| record
+	;
+
+record:	header fields CLOSE
+	{ printf("\t}\n"); }
+	;
+
+header: PV WORD TYPE WORD WORD
+	{ printf("\trecord(%s,\"%s\") {\n",$4,$2); }
+	| PV WORD TYPE WORD
+	{ printf("\trecord(%s,\"%s\") {\n",$4,$2); }
+
+fields: field
+	| fields field
+	;
+
+field:	a_field words
+	{
+		printf("\t\tfield(%s,\"%s\")\n",$<Str>1,Field_val);
+	}
+	;
+
+a_field: FIELD
+	{
+		Field_val[0]='\0';
+		$<Str>$ = $1;
+	}
+	;
+
+words: 
+	| words a_word
+	;
+
+a_word: WORD
+	{
+		char *p;
+
+		if((p=strstr($1,".PP.MS"))) {p[0]=' ';p[3]=' '; }
+		else if((p=strstr($1,".PP.NMS"))) {p[0]=' ';p[3]=' ';}
+		else if((p=strstr($1,".NPP.MS"))) {p[0]=' ';p[4]=' ';}
+		else if((p=strstr($1,".NPP.NMS"))) {p[0]=' ';p[4]=' ';}
+		else { if(strlen(Field_val)>0) strcat(Field_val," "); }
+
+		strcat(Field_val,$1);
+	}
+	;
+
+%%
+ 
+#include "sf2dbLex.c"
+ 
+yyerror(str)
+char  *str;
+{ fprintf(stderr,"Error line %d : %s\n",line_num, yytext); }
+ 
+/*-----------------------main routine-----------------------*/
+main(argc, argv)
+int argc;
+char **argv;
+{
+FILE *fdin;
+
+	/* remember that the database name is not retrieved from the .report
+	   file, it is specified as a command line argument */
+
+	if( argc < 2 )
+	{
+		printf("Usage: %s database_name < dct_report_file > new_gdct_db_file.db\n",
+			argv[0]);
+		printf("\n\twhere\n\tdatabase_name: the name you wish to give the .database file\n");
+		printf("\tdct_report_file: the old dct short form report file\n");
+		printf("\tnew_gdct_db_file.db: the new gdct db file name, this file is used as \n\t\tinput to the gdct.\n");
+		exit(0);
+	}
+
+#ifdef PUKE_HEAD
+	printf("database(x) { nowhere() {\n");
+#endif
+	/* yyreset(); */
+	yyparse();
+#ifdef PUKE_HEAD
+	printf("}}\n");
+#endif
+
+
+/*	fdin = freopen("tester.report", "r+", stdin);
+	yyreset();
+	yyparse(); */
+}
+