diff --git a/CHANGES b/CHANGES
new file mode 100644
index 00000000..29c1503a
--- /dev/null
+++ b/CHANGES
@@ -0,0 +1,24 @@
+1 May 2003
+
+RELEASE 0.0.2
+
+1 June 2003
+
+RELEASE 0.0.4
+
+21 June 2003
++ Fix error where we could allocate a used sector!  ARGH!
+
+RELEASE 0.0.5
+
+18 July 2004
++ Minor bug fix
+
+RELEASE 0.0.6
+
+19 August 2007
++ Fix so compiles with newish gcc
++ Update asoft_detoken to match documentation found online
++ Create tokenize_asoft
+
+RELEASE 0.0.7
diff --git a/Changelog b/Changelog
deleted file mode 100644
index 423be330..00000000
--- a/Changelog
+++ /dev/null
@@ -1,2 +0,0 @@
-21 June 2003
-+ Fix error where we could allocate a used sector!  ARGH!
diff --git a/Makefile b/Makefile
index 83f818bc..80fe49a6 100644
--- a/Makefile
+++ b/Makefile
@@ -2,7 +2,7 @@ CC = gcc
 C_FLAGS = -O2 -Wall
 L_FLAGS =
 
-all:	dos33 asoft_detoken mkdos33fs make_b
+all:	dos33 asoft_detoken mkdos33fs make_b tokenize_asoft
 
 
 
@@ -12,7 +12,11 @@ asoft_detoken:		   asoft_detoken.o
 asoft_detoken.o:	   asoft_detoken.c
 			   $(CC) $(C_FLAGS) -c asoft_detoken.c
 			 
+tokenize_asoft:		   tokenize_asoft.o
+			   $(CC) $(L_FLAGS) -o tokenize_asoft tokenize_asoft.o
 			   
+tokenize_asoft.o:	   tokenize_asoft.c
+			   $(CC) $(C_FLAGS) -c tokenize_asoft.c
 
 dos33:	dos33.o
 		$(CC) $(L_FLAGS) -o dos33 dos33.o
@@ -35,7 +39,7 @@ mkdos33fs.o:	mkdos33fs.c dos33.h
 
 
 install:	
-		cp dos33 asoft_detoken mkdos33fs make_b /usr/local/bin
+		cp dos33 asoft_detoken mkdos33fs tokenize_asoft make_b /usr/local/bin
 
 clean:		
-		rm -f *~ *.o asoft_detoken dos33 make_b mkdos33fs
+		rm -f *~ *.o asoft_detoken dos33 make_b mkdos33fs tokenize_asoft
diff --git a/README b/README
new file mode 100644
index 00000000..ca9b3d83
--- /dev/null
+++ b/README
@@ -0,0 +1,19 @@
+dos33fsprogs
+
+    by Vince Weaver <vince _at_ deater.net>
+    
+
+These are some tools for manipulating Apple II disk files that
+I've written over the years while doing Apple 2 hacking.
+
+make_b : take a machine language blob and give it the size/offset
+         header needed to BLOAD it from DOS3.3
+	 
+asoft_detoken: takes an applesoft file obtained with dos33
+               and converts it to an ASCII text file
+	       
+tokenize_asoft: takes an ASCII text file and converts it
+                into a tokenized applesoft file
+		
+dos33:		a tool for manipulating dos33 .dsk images
+
diff --git a/TODO b/TODO
index d1bef099..d67ba3b6 100644
--- a/TODO
+++ b/TODO
@@ -4,3 +4,5 @@ defrag utility
 test-suite
 
 test mkdosfs copy over DOS functionality
+
+find out why dos33 SAVE puts extra linefeed at end of basic prog?
diff --git a/asoft_detoken.c b/asoft_detoken.c
index 4f7cce7a..3509e4e4 100644
--- a/asoft_detoken.c
+++ b/asoft_detoken.c
@@ -22,130 +22,36 @@ char applesoft_tokens[][8]={
 /* F8 */ "","","","","","(","(","("
 };
 
-/*
-Integer Basic
+#if 0
+/* Integer Basic */
 
-         $03: :
-         $04: LOAD
-         $05: SAVE
-         $07: RUN
-         $09: DEL
-         $0A: ,
-         $0B: NEW
-         $0C: CLR
-         $0D: AUTO
-         $0F: MAN
-         $10: HIMEM:
-         $11: LOMEM:
-         $12: +
-         $13: -
-         $14: *
-         $15: /
-         $16: =
-         $17: #
-         $18: >=
-         $19: >
-         $1A: <=
-         $1B: <>
-         $1C: <
-         $1D:  AND
-         $1E:  OR
-         $1F:  MOD
-         $20: ^
-         $22: (
-         $23: ,
-         $24,
-         $25:  THEN
-         $26,
-         $27: ,
-         $28, $29: \"
-         $2A: (
-         $2D: (
-         $2E:  PEEK
-         $2F: RND
-         $30: SGN
-         $31: ABS
-         $32: PDL
-         $34: (
-         $35: +
-         $36: -
-         $37: NOT
-         $38: (
-         $39: =
-         $3A: #
-         $3B: LEN (
-         $3C: ASC (
-         $3D: SCRN (
-         $3E: ,
-         $3F:  (
-         $40: $
-         $42: (
-         $43,
-         $44: ,
-         $45,
-         $46,
-         $47: ;
-         $48,
-         $49: ,
-         $4A: ,
-         $4B: TEXT
-         $4C: GR
-         $4D: CALL
-         $4E,
-         $4F: DIM
-         $50: TAB
-         $51: END
-         $52,  $53,  $54: INPUT
-         $55: FOR
-         $56:  =
-         $57:  TO
-         $58:  STEP
-         $59: NEXT
-         $5A: ,
-         $5B: RETURN
-         $5C: GOSUB
-         $5D: REM
-         $5E: LET
-         $5F: GOTO
-         $60: IF
-         $61,
-         $62: PRINT
-         $63: PRINT
-         $64:    POKE
-         $65: ,
-         $66: COLOR=
-         $67: PLOT
-         $68: ,
-         $69: HLIN
-         $6A: ,
-         $6B:  AT
-         $6C: VLIN
-         $6D: ,
-         $6E:  AT
-         $6F: VTAB
-         $70,
-         $71:  =
-         $72: )
-         $74: LIST
-         $75: ,
-         $77: POP
-         $79: NO DSP
-         $7A: NO TRACE
-         $7B,
-         $7C: DSP
-         $7D: TRACE
-         $7E: PR #
-         $7F: IN #
-
-*/
+char integer_tokens[][8]={   
+/* 00 */ "","","",":","LOAD","SAVE","","RUN",
+/* 08 */ "","DEL",",","NEW","CLR","AUTO","","MAN",
+/* 10 */ "HIMEM:","LOMEM:","+","-","*","/","=","#",
+/* 18 */ ">=",">","<=","<>","<"," AND"," OR"," MOD",
+/* 20 */ "^","","(",",",""," THEN","",",",
+/* 28 */ "\"","\"","(","","","("," PEEK","RND",
+/* 30 */ "SGN","ABS","PDL","","(","+","-","NOT",
+/* 38 */ "(","=","LEN (","ASC (","SCRN (",","," (",
+/* 40 */ "$","(","",",","","",";",
+/* 48 */ "",",",",","TEXT","GR","CALL","","DIM",
+/* 50 */ "TAB","END","","","INPUT","FOR","=","TO",
+/* 58 */ " STEP","NEXT",",","RETURN","GOSUB","REM","LET","GOTO",
+/* 60 */ "IF","","PRINT","PRINT","   POKE",",","COLOR=","PLOT",
+/* 68 */ ",","HLIN",","," AT","VLIN",","," AT","VTAB",
+/* 70 */ "","=",")","","LIST",",","","POP",
+/* 78 */ "","NO DSP","NO TRACE","","DSP","TRACE","PR #","IN #"
+};
+#endif 
 		
 		
 int main(int argc, char **argv) {
    
    int ch1,i;
-   unsigned char size1,size2;
-   unsigned char line1,line2;
-   unsigned char eight,line_length;
+   int size1,size2;
+   int line1,line2;
+   int link1,link2,link;
    
    
        /* read size, first two bytes */
@@ -153,17 +59,24 @@ int main(int argc, char **argv) {
    size2=fgetc(stdin);
    
    while(!feof(stdin)) {
-  
+        
+      /* link points to the next line */
+      /* assumes asoft program starts at address $801 */
+      link1=fgetc(stdin);
+      link2=fgetc(stdin);
+      link=(link1<<8)|link2;
+      /* link==0 indicates EOF */
+      if (link==0) goto the_end;
       
-      line_length=fgetc(stdin);
-      eight=fgetc(stdin);   /* sometimes 8, sometimes 9? */
-      if (eight==0) goto the_end;
+      /* line number is little endian 16-bit value */
       line1=fgetc(stdin);
       line2=fgetc(stdin);
       if (feof(stdin)) goto the_end;
-      printf("%4d ",(((int)line2)<<8)+line1);
-	
+      printf("%4d ",((line2)<<8)+line1);
+      
+      /* repeat until EOL character (0) */
       while( (ch1=fgetc(stdin))!=0 ) {
+	     /* if > 0x80 it's a token */
 	  if (ch1>=0x80) {
 	     fputc(' ',stdout);
 	     for(i=0;i<strlen(applesoft_tokens[ch1-0x80]);i++) {
@@ -171,7 +84,11 @@ int main(int argc, char **argv) {
 	     }
 	     fputc(' ',stdout);
 	  }
-          else fputc(ch1,stdout);
+            /* otherwise it is an ascii char */
+	  else {
+	     fputc(ch1,stdout);
+	  }
+	 
       }
       printf("\n");
    }
diff --git a/dos33.c b/dos33.c
index c70e6fdb..259863d5 100644
--- a/dos33.c
+++ b/dos33.c
@@ -843,7 +843,7 @@ repeat_tsl:
        ts_s=tslist[TSL_NEXT_SECTOR];
    
        if (!((ts_s==0) && (ts_t==0))) goto repeat_tsl;
-continue_dump:
+continue_dump:;
     }
     
     catalog_t=buffer[CATALOG_NEXT_T];
diff --git a/tests/GR b/tests/GR
new file mode 100644
index 00000000..424281ff
Binary files /dev/null and b/tests/GR differ
diff --git a/tests/SINCOS b/tests/SINCOS
new file mode 100644
index 00000000..c467de3a
Binary files /dev/null and b/tests/SINCOS differ
diff --git a/YOU.LOGO b/tests/YOU.LOGO
similarity index 100%
rename from YOU.LOGO
rename to tests/YOU.LOGO
diff --git a/tokenize_asoft.c b/tokenize_asoft.c
new file mode 100644
index 00000000..585cc4e7
--- /dev/null
+++ b/tokenize_asoft.c
@@ -0,0 +1,220 @@
+#include <stdio.h>
+#include <string.h> /* strlen() */
+#include <stdlib.h> /* exit() */
+
+/* TODO */
+/* match lowecase tokens as well as upper case ones */
+
+
+
+/* Info from http://docs.info.apple.com/article.html?coll=ap&artnum=57 */
+
+/* In memory, applesoft file starts at address $801        */
+/* format is <LINE><LINE><LINE>$00$00                      */
+/* Where <LINE> is:                                        */
+/*    2 bytes (little endian) of LINK indicating addy of next line */ 
+/*    2 bytes (little endian) giving the line number       */
+/*    a series of bytes either ASCII or tokens (see below) */
+/*    a $0 char indicating end of line                     */
+
+
+   /* Starting at 0x80 */
+char applesoft_tokens[][8]={
+   
+/* 80 */ "END","FOR","NEXT","DATA","INPUT","DEL","DIM","READ",
+/* 88 */ "GR","TEXT","PR #","IN #","CALL","PLOT","HLIN","VLIN",
+/* 90 */ "HGR2","HGR","HCOLOR=","HPLOT","DRAW","XDRAW","HTAB","HOME",
+/* 98 */ "ROT=","SCALE=","SHLOAD","TRACE","NOTRACE","NORMAL","INVERSE","FLASH",
+/* A0 */ "COLOR=","POP","VTAB ","HIMEM:","LOMEM:","ONERR","RESUME","RECALL",
+/* A8 */ "STORE","SPEED=","LET","GOTO","RUN","IF","RESTORE","&",
+/* B0 */ "GOSUB","RETURN","REM","STOP","ON","WAIT","LOAD","SAVE",
+/* B8 */ "DEF FN","POKE","PRINT","CONT","LIST","CLEAR","GET","NEW",
+/* C0 */ "TAB","TO","FN","SPC(","THEN","AT","NOT","STEP",
+/* C8 */ "+","-","*","/","^","AND","OR",">",
+/* D0 */ "=","<","SGN","INT","ABS","USR","FRE","SCRN (",
+/* D8 */ "PDL","POS","SQR","RND","LOG","EXP","COS","SIN",
+/* E0 */ "TAN","ATN","PEEK","LEN","STR$","VAL","ASC","CHR$",
+/* E8 */ "LEFT$","RIGHT$","MID$","","","","","",
+/* F0 */ "","","","","","","","",
+/* F8 */ "","","","","","(","(","("
+};
+
+#if 0
+/* Integer Basic */
+
+char integer_tokens[][8]={   
+/* 00 */ "","","",":","LOAD","SAVE","","RUN",
+/* 08 */ "","DEL",",","NEW","CLR","AUTO","","MAN",
+/* 10 */ "HIMEM:","LOMEM:","+","-","*","/","=","#",
+/* 18 */ ">=",">","<=","<>","<"," AND"," OR"," MOD",
+/* 20 */ "^","","(",",",""," THEN","",",",
+/* 28 */ "\"","\"","(","","","("," PEEK","RND",
+/* 30 */ "SGN","ABS","PDL","","(","+","-","NOT",
+/* 38 */ "(","=","LEN (","ASC (","SCRN (",","," (",
+/* 40 */ "$","(","",",","","",";",
+/* 48 */ "",",",",","TEXT","GR","CALL","","DIM",
+/* 50 */ "TAB","END","","","INPUT","FOR","=","TO",
+/* 58 */ " STEP","NEXT",",","RETURN","GOSUB","REM","LET","GOTO",
+/* 60 */ "IF","","PRINT","PRINT","   POKE",",","COLOR=","PLOT",
+/* 68 */ ",","HLIN",","," AT","VLIN",","," AT","VTAB",
+/* 70 */ "","=",")","","LIST",",","","POP",
+/* 78 */ "","NO DSP","NO TRACE","","DSP","TRACE","PR #","IN #"
+};
+#endif 
+
+#define LOW(_x)  ((_x)&0xff) 
+#define HIGH(_x) (((_x)>>8)&0xff)
+#define MAXSIZE 65535
+
+/* File cannot be longer than 64k */
+unsigned char output[MAXSIZE+1];
+
+char *line_ptr;
+int line=0;
+char input_line[BUFSIZ];
+
+static void show_problem(char *line_ptr) {
+   int offset,i;
+   
+   offset=(int)(line_ptr-input_line);
+   fprintf(stderr,"%s",input_line);
+   for(i=0;i<offset;i++) fputc(' ',stderr);
+   fprintf(stderr,"^\n");
+}
+
+   
+
+static int getnum(void) {
+   int num=0;
+   
+   /* skip any whitespace */
+   while((*line_ptr<=' ') && (*line_ptr!=0)) line_ptr++;
+   
+   while (*line_ptr>' ') {
+      if ((*line_ptr<'0')||(*line_ptr>'9')) {
+	 fprintf(stderr,"Invalid line number line %d\n",line);
+	 show_problem(line_ptr);
+	 exit(-1);
+      }
+      num*=10;
+      num+=(*line_ptr)-'0';
+      line_ptr++;
+   }
+
+   if (!line_ptr) {
+      fprintf(stderr,"Missing line number line %d\n",line);
+      exit(-1);
+   }
+      
+   return num;	
+}
+
+static int in_quotes;
+
+static int find_token() {
+   
+   int ch,i;
+   
+   ch=*line_ptr;
+   /* don't skip whitespace in quotes */
+   if (!in_quotes) {	
+      while(ch<=' ') {	
+         if ((ch=='\n') || (ch=='\r') || (ch=='\0')) {
+            return 0;
+         }
+         line_ptr++;
+         ch=*line_ptr;
+      }
+   }
+   
+   
+      /* toggle quotes mode */
+   if (ch=='\"') in_quotes=!in_quotes;
+	
+      /* don't tokenize if in quotes */
+   if (!in_quotes) {
+	
+  // fprintf(stderr,"%s",line_ptr);
+   for(i=0;i<107;i++) {
+      if (!strncmp(line_ptr,applesoft_tokens[i],strlen(applesoft_tokens[i]))) {
+//	 fprintf(stderr,"Found token %x (%s) %d\n",0x80+i,applesoft_tokens[i],i);
+	 line_ptr+=strlen(applesoft_tokens[i]);
+	 return 0x80+i;
+      }
+//      fprintf(stderr,"%s ",applesoft_tokens[i]);
+   }
+   }
+   
+//   fprintf(stderr,"\n");
+      /* not a token, just ascii */
+   line_ptr++;
+   return ch;
+}
+
+static void check_oflo(int size) {
+   if (size>MAXSIZE) {
+      fprintf(stderr,"Output file too big!\n");
+      exit(-1);
+   }   
+}
+
+	
+   
+
+int main(int argc, char **argv) {
+   
+   int offset=2,i;
+
+   int linenum,link_offset;
+   int link_value=0x801; /* start of applesoft program */
+   int token;
+
+   while(1) {
+         /* get line from input file */
+      line_ptr=fgets(input_line,BUFSIZ,stdin);
+      line++;
+      if (line_ptr==NULL) break;
+      linenum=getnum();
+      if (linenum>65535) {
+	 printf("Invalid line number %d\n",linenum);
+	 exit(-1);
+      }
+      link_offset=offset;
+      check_oflo(offset+4);
+      output[offset+2]=LOW(linenum);
+      output[offset+3]=HIGH(linenum);
+      offset+=4;
+      
+      
+      while(1) {
+	 token=find_token();
+	 output[offset]=token;
+	 offset++;
+	 check_oflo(offset);
+	 if (!token) break;
+      }
+      
+	 
+      /* 2 bytes is to ignore size from beginning of file */
+      link_value=0x801+(offset-2);
+      
+      /* point link value to next line */
+      check_oflo(offset+2);
+      output[link_offset]=LOW(link_value);
+      output[link_offset+1]=HIGH(link_value);
+   }
+   /* set last link field to $00 $00 which indicates EOF */
+   check_oflo(offset+2);
+   output[offset]='\0';
+   output[offset+1]='\0';
+   offset+=2;
+   
+     /* Set filesize */
+     /* -1 to match observed values */
+   output[0]=LOW(offset-1);
+   output[1]=HIGH(offset-1);
+   /* output our file */
+   for(i=0;i<offset;i++) putchar(output[i]);
+   
+   return 0;
+}