diff --git a/math32.tal b/math32.tal
index cce6606..3f412a2 100644
--- a/math32.tal
+++ b/math32.tal
@@ -110,6 +110,16 @@
 @complement32 ( x** -> ~x** )
     COMPLEMENT32 RTN
 
+( temporary registers )
+( used by all operations except div32 )
+@sh [
+  &r $1
+  &x0 $1 &x1 $1 &x2 $1 &x3 $1
+  &y0 $1 &y1 $1 &y2 $1 &y3 $1
+  &z0 $1 &z1 $1 &z2 $1 &z3 $1
+  &a0 $1 &a1 $1 &a2 $2
+]
+
 ( bit shifting )
 
 @right-shift ( x** n^ -> x<<n )
@@ -121,46 +131,42 @@
 
 ( shift right by 0-7 bits )
 @right-shift0 ( x** n^ -> x<<n )
-    #0000 ,&z0 STR2 #0000 ,&z2 STR2
-    ,&r STR ( n -> r )
-    ,&r LDR SFT ,&z3 STR ( write z3 )
-    #00 ,&r LDR SFT2 ,&z2 LDR2 ORA2 ,&z2 STR2 ( write z2,z3 )
-    #00 ,&r LDR SFT2 ,&z1 LDR2 ORA2 ,&z1 STR2 ( write z1,z2 )
-    #00 ,&r LDR SFT2 ,&z0 LDR2 ORA2 ,&z0 STR2 ( write z0,z1 )
-    ,&z0 LDR2 ,&z2 LDR2
+    #0000 ;sh/z0 STA2 #0000 ;sh/z2 STA2
+    ;sh/r STA ( n -> r )
+    ;sh/r LDA SFT ;sh/z3 STA ( write z3 )
+    #00 ;sh/r LDA SFT2 ;sh/z2 LDA2 ORA2 ;sh/z2 STA2 ( write z2,z3 )
+    #00 ;sh/r LDA SFT2 ;sh/z1 LDA2 ORA2 ;sh/z1 STA2 ( write z1,z2 )
+    #00 ;sh/r LDA SFT2 ;sh/z0 LDA2 ORA2 ;sh/z0 STA2 ( write z0,z1 )
+    ;sh/z0 LDA2 ;sh/z2 LDA2
     RTN
-[ &r $1 &z0 $1 &z1 $1 &z2 $1 &z3 $1 ]
 
 ( shift right by 8-15 bits )
 @right-shift1 ( x** n^ -> x<<n )
-    #00 ,&z1 STR #0000 ,&z2 STR2
-    #08 SUB ,&r STR ( n -> r )
+    #00 ;sh/z1 STA #0000 ;sh/z2 STA2
+    #08 SUB ;sh/r STA ( n -> r )
     POP
-    ,&r LDR SFT ,&z3 STR ( write z3 )
-    #00 ,&r LDR SFT2 ,&z2 LDR2 ORA2 ,&z2 STR2 ( write z2,z3 )
-    #00 ,&r LDR SFT2 ,&z1 LDR2 ORA2 ,&z1 STR2 ( write z1,z2 )
-    #00 ,&z1 LDR ,&z2 LDR2
+    ;sh/r LDA SFT ;sh/z3 STA ( write z3 )
+    #00 ;sh/r LDA SFT2 ;sh/z2 LDA2 ORA2 ;sh/z2 STA2 ( write z2,z3 )
+    #00 ;sh/r LDA SFT2 ;sh/z1 LDA2 ORA2 ;sh/z1 STA2 ( write z1,z2 )
+    #00 ;sh/z1 LDA ;sh/z2 LDA2
     RTN
-[ &r $1 &z1 $1 &z2 $1 &z3 $1 ]
 
 ( shift right by 16-23 bits )
 @right-shift2 ( x** n^ -> x<<n )
-    #0000 ,&z2 STR2
-    #10 SUB ,&r STR ( n -> r )
+    #0000 ;sh/z2 STA2
+    #10 SUB ;sh/r STA ( n -> r )
     POP2
-    ,&r LDR SFT ,&z3 STR ( write z3 )
-    #00 ,&r LDR SFT2 ,&z2 LDR2 ORA2 ,&z2 STR2 ( write z2,z3 )
-    #0000 ,&z2 LDR2
+    ;sh/r LDA SFT ;sh/z3 STA ( write z3 )
+    #00 ;sh/r LDA SFT2 ;sh/z2 LDA2 ORA2 ;sh/z2 STA2 ( write z2,z3 )
+    #0000 ;sh/z2 LDA2
     RTN
-[ &r $1 &z2 $1 &z3 $1 ]
 
 ( shift right by 16-23 bits )
 @right-shift3 ( x** n^ -> x<<n )
-    #18 SUB ,&r STR ( n -> r )
+    #18 SUB ;sh/r STA ( n -> r )
     POP2 POP #00 SWP #0000 SWP2 ( 00 00 00 x0 )
-    ,&r LDR SFT
+    ;sh/r LDA SFT
     RTN
-[ &r $1 ]
 
 @left-shift ( x** n^ -> x<<n )
     DUP #08 LTH ;left-shift0 JCN2 ( x n )
@@ -171,57 +177,54 @@
 
 ( shift left by 0-7 bits )
 @left-shift0 ( x** n^ -> x<<n )
-    #0000 ,&z0 STR2 #0000 ,&z2 STR2
-    #40 SFT ,&r STR ( n<<4 -> r )
+    #0000 ;sh/z0 STA2 #0000 ;sh/z2 STA2
+    #40 SFT ;sh/r STA ( n<<4 -> r )
     SWP SWP2 SWP ( x3 x2 x1 x0 )
 
-    ,&r LDR SFT ,&z0 STR ( x3 x2 x1 )
+    ;sh/r LDA SFT ;sh/z0 STA ( x3 x2 x1 )
 
-    #00 SWP ,&r LDR SFT2 ( x3 x2 00x1<<r )
-    ,&z0 LDR2 ORA2 ,&z0 STR2 ( x3 x2 )
+    #00 SWP ;sh/r LDA SFT2 ( x3 x2 00x1<<r )
+    ;sh/z0 LDA2 ORA2 ;sh/z0 STA2 ( x3 x2 )
 
-    #00 SWP ,&r LDR SFT2 ( x3 00x2<<r )
-    ,&z1 LDR2 ORA2 ,&z1 STR2 ( x3 )
+    #00 SWP ;sh/r LDA SFT2 ( x3 00x2<<r )
+    ;sh/z1 LDA2 ORA2 ;sh/z1 STA2 ( x3 )
 
-    #00 SWP ,&r LDR SFT2 ( 00x3<<r )
-    ,&z2 LDR2 ORA2 ,&z2 STR2 ( )
+    #00 SWP ;sh/r LDA SFT2 ( 00x3<<r )
+    ;sh/z2 LDA2 ORA2 ;sh/z2 STA2 ( )
 
-    ,&z0 LDR2 ,&z2 LDR2
+    ;sh/z0 LDA2 ;sh/z2 LDA2
     RTN
-[ &r $1 &z0 $1 &z1 $1 &z2 $2 ]
 
 ( shift left by 8-15 bits )
 @left-shift1 ( x** n^ -> x<<n )
-    #0000 ,&z0 STR2 #00 ,&z2 STR
-    #08 SUB #40 SFT ,&r STR ( n<<4 -> r )
+    #0000 ;sh/z0 STA2 #00 ;sh/z2 STA
+    #08 SUB #40 SFT ;sh/r STA ( n<<4 -> r )
     SWP SWP2 SWP POP ( x3 x2 x1 )
 
-    ,&r LDR SFT ,&z0 STR ( x3 x2 )
+    ;sh/r LDA SFT ;sh/z0 STA ( x3 x2 )
 
-    #00 SWP ,&r LDR SFT2 ( x3 00x2<<r )
-    ,&z0 LDR2 ORA2 ,&z0 STR2 ( x3 )
+    #00 SWP ;sh/r LDA SFT2 ( x3 00x2<<r )
+    ;sh/z0 LDA2 ORA2 ;sh/z0 STA2 ( x3 )
 
-    #00 SWP ,&r LDR SFT2 ( 00x3<<r )
-    ,&z1 LDR2 ORA2 ,&z1 STR2 ( )
+    #00 SWP ;sh/r LDA SFT2 ( 00x3<<r )
+    ;sh/z1 LDA2 ORA2 ;sh/z1 STA2 ( )
 
-    ,&z0 LDR2 ,&z2 LDR #00
+    ;sh/z0 LDA2 ;sh/z2 LDA #00
     RTN
-[ &r $1 &z0 $1 &z1 $1 &z2 $1 ]
 
 ( shift left by 16-23 bits )
 @left-shift2 ( x** n^ -> x<<n )
-    #0000 ,&z0 STR2
-    #10 SUB #40 SFT ,&r STR ( n<<4 -> r )
+    #0000 ;sh/z0 STA2
+    #10 SUB #40 SFT ;sh/r STA ( n<<4 -> r )
     SWP2 POP2 SWP ( x3 x2 )
 
-    ,&r LDR SFT ,&z0 STR ( x3 )
+    ;sh/r LDA SFT ;sh/z0 STA ( x3 )
 
-    #00 SWP ,&r LDR SFT2 ( x3<<r )
-    ,&z0 LDR2 ORA2 ,&z0 STR2 (  )
+    #00 SWP ;sh/r LDA SFT2 ( x3<<r )
+    ;sh/z0 LDA2 ORA2 ;sh/z0 STA2 (  )
 
-    ,&z0 LDR2 #0000
+    ;sh/z0 LDA2 #0000
     RTN
-[ &r $1 &z0 $2 ]
 
 ( shift left by 24-31 bits )
 @left-shift3 ( x** n^ -> x<<n )
@@ -233,32 +236,28 @@
 ( arithmetic )
 
 @add32 ( xhi* xlo* yhi* ylo* -> zhi* zlo* )
-    ,&y2 STR2 ,&y0 STR2 ( save ylo, yhi )
-    ,&x2 STR2 ,&x0 STR2 ( save xlo, xhi )
-    #0000 #0000 ,&z0 STR2 ,&z2 STR2 ( reset zhi, zlo )
+    ;sh/y2 STA2 ;sh/y0 STA2 ( save ylo, yhi )
+    ;sh/x2 STA2 ;sh/x0 STA2 ( save xlo, xhi )
+    #0000 #0000 ;sh/z0 STA2 ;sh/z2 STA2 ( reset zhi, zlo )
 
     ( x3 + y3 => z2z3 )
-    #00 ,&x3 LDR #00 ,&y3 LDR ADD2 ,&z2 STR2
+    #00 ;sh/x3 LDA #00 ;sh/y3 LDA ADD2 ;sh/z2 STA2
 
     ( x2 + y2 + z2 => z1z2 )
-    #00 ,&x2 LDR ,&z1 LDR2 ADD2 ,&z1 STR2
-    #00 ,&y2 LDR ,&z1 LDR2 ADD2 ,&z1 STR2
+    #00 ;sh/x2 LDA ;sh/z1 LDA2 ADD2 ;sh/z1 STA2
+    #00 ;sh/y2 LDA ;sh/z1 LDA2 ADD2 ;sh/z1 STA2
 
     ( x1 + y1 + z1 => z0z1 )
-    #00 ,&x1 LDR ,&z0 LDR2 ADD2 ,&z0 STR2
-    #00 ,&y1 LDR ,&z0 LDR2 ADD2 ,&z0 STR2
+    #00 ;sh/x1 LDA ;sh/z0 LDA2 ADD2 ;sh/z0 STA2
+    #00 ;sh/y1 LDA ;sh/z0 LDA2 ADD2 ;sh/z0 STA2
 
     ( x0 + y0 + z0 => z0 )
-    ,&x0 LDR ,&z0 LDR ADD ,&z0 STR
-    ,&y0 LDR ,&z0 LDR ADD ,&z0 STR
+    ;sh/x0 LDA ;sh/z0 LDA ADD ;sh/z0 STA
+    ;sh/y0 LDA ;sh/z0 LDA ADD ;sh/z0 STA
 
     ( load zhi,zlo )
-    ,&z0 LDR2 ,&z2 LDR2
+    ;sh/z0 LDA2 ;sh/z2 LDA2
     RTN
-( registers for add32 )
-[ &x0 $1 &x1 $1 &x2 $1 &x3 $1 ]
-[ &y0 $1 &y1 $1 &y2 $1 &y3 $1 ]
-[ &z0 $1 &z1 $1 &z2 $2 ]
 
 @negate32 ( x** -> -x** )
     COMPLEMENT32
@@ -273,48 +272,44 @@
     ;negate32 JSR2 ;add32 JSR2 RTN
 
 @mul16 ( x* y* -> z** )
-    ,&y1 STR ,&y0 STR ( save ylo, yhi )
-    ,&x1 STR ,&x0 STR ( save xlo, xhi )
-    #0000 #00 ,&z0 STR2 ,&z2 STR ( reset z0,z1,z2 )
-    #0000 #00 ,&a0 STR2 ,&a2 STR ( reset a0,a1,a2 )
+    ;sh/y1 STA ;sh/y0 STA ( save ylo, yhi )
+    ;sh/x1 STA ;sh/x0 STA ( save xlo, xhi )
+    #0000 #00 ;sh/z1 STA2 ;sh/z3 STA ( reset z1,z2,z3 )
+    #0000 #00 ;sh/a0 STA2 ;sh/a2 STA ( reset a0,a1,a2 )
 
     ( x1 * y1 => z1z2 )
-    #00 ,&x1 LDR #00 ,&y1 LDR MUL2 ,&z1 STR2
+    #00 ;sh/x1 LDA #00 ;sh/y1 LDA MUL2 ;sh/z2 STA2
 
     ( x0 * y1 => z0z1 )
-    #00 ,&x0 LDR #00 ,&y1 LDR MUL2 ,&z0 LDR2 ADD2 ,&z0 STR2
+    #00 ;sh/x0 LDA #00 ;sh/y1 LDA MUL2 ;sh/z1 LDA2 ADD2 ;sh/z1 STA2
 
     ( x1 * y0 => a1a2 )
-    #00 ,&x1 LDR #00 ,&y0 LDR MUL2 ,&a1 STR2
+    #00 ;sh/x1 LDA #00 ;sh/y0 LDA MUL2 ;sh/a1 STA2
 
     ( x0 * y0 => a0a1 )
-    #00 ,&x0 LDR #00 ,&y0 LDR MUL2 ,&a0 LDR2 ADD2 ,&a0 STR2
+    #00 ;sh/x0 LDA #00 ;sh/y0 LDA MUL2 ;sh/a0 LDA2 ADD2 ;sh/a0 STA2
 
     ( add z and a<<8 )
-    #00 ,&z0 LDR2 ,&z2 LDR
-    ,&a0 LDR2 ,&a2 LDR #00
+    #00 ;sh/z1 LDA2 ;sh/z3 LDA
+    ;sh/a0 LDA2 ;sh/a2 LDA #00
     ;add32 JSR2
     RTN
-[ &x0 $1 &x1 $1 ]
-[ &y0 $1 &y1 $1 ]
-[ &z0 $1 &z1 $1 &z2 $1 ]
-[ &a0 $1 &a1 $1 &a2 $1 ]
 
 @mul32 ( x** y** -> z** )
-    ,&y1 STR2 ,&y0 STR2 ( save ylo, yhi )
-    ,&x1 STR2 ,&x0 STR2 ( save xlo, xhi )
-    ,&y1 LDR2 ,&x1 LDR2 ;mul16 JSR2 ( [x1*y1] )
-    ,&z1 STR2 ,&z0 STR2 ( sum = x1*y1, save zlo, zhi )
+    ,&y2 STR2 ,&y0 STR2 ( save ylo, yhi )
+    ,&x2 STR2 ,&x0 STR2 ( save xlo, xhi )
+    ,&y2 LDR2 ,&x2 LDR2 ;mul16 JSR2 ( [x2*y2] )
+    ,&z2 STR2 ,&z0 STR2 ( sum = x2*y2, save zlo, zhi )
 
-    ,&y1 LDR2 ,&x0 LDR2 MUL2 ( [x0*y1]<<16 )
-    ,&y0 LDR2 ,&x1 LDR2 MUL2 ( [x1*y0]<<16 )
+    ,&y2 LDR2 ,&x0 LDR2 MUL2 ( [x0*y2]<<16 )
+    ,&y0 LDR2 ,&x2 LDR2 MUL2 ( [x2*y0]<<16 )
     ( [x0*y0]<<32 will completely overflow )
-    ADD2 ,&z0 LDR2 ADD2 ( sum += x0*y1<<16 + x1*y0<<16 )
-    ,&z1 LDR2
+    ADD2 ,&z0 LDR2 ADD2 ( sum += x0*y2<<16 + x2*y0<<16 )
+    ,&z2 LDR2
     RTN
-[ &x0 $2 &x1 $2 ]
-[ &y0 $2 &y1 $2 ]
-[ &z0 $2 &z1 $2 ]
+[ &x0 $2 &x2 $2 ]
+[ &y0 $2 &y2 $2 ]
+[ &z0 $2 &z2 $2 ]
 
 @div32 ( x** y** -> q** )
     ( store y and x for repeated use )