To: J3                                                     J3/25-171
From: Van Snyder
Subject: Edits if 25-170 "NORM2 needs a MASK argument" passes
Date: 2025-September-26
Reference: 25-007 25-170


Edits
-----

[384:+8 Table 16.1 Standard generic intrinsic procedure summary (cont.)]
Replace the eighth entry on the page by

"NORM2 (X [,MASK] ) or NORM2 (X, DIM [,MASK]  T  $L_2$ norm of an array"

[457:12 16.9.153 NORM2 (X) or NORM2 (X,DIM)]
Replace the title by

"NORM2 (X [,MASK] ) or NORM2 (X, DIM [,MASK])

[457:17+ 16.9.153p3 Arguments] Insert an argument description

"MASK(optional)  shall be of type logical and shall be conformable
                 with X." {with "conformable" linked to its definition}.

[457:25-27 16.9.153p5 Result Value] Replace "Case(ii) ..." with two
cases

"Case(ii):  The result of NORM2 (X, MASK=MASK) has a value equal to a
            processor-dependent approximation to the $L_2$ norm of the
            elements of X corresponding to true elements of MASK or has
            the value zero if there are no true elements.

"Case(iii): If X has rank one, NORM2 (X, DIM=DIM [,MASK=MASK] ) has a
            value equal to that of NORM2 (X, MASK=MASK). Otherwise the
            value of elements ($s_1$, $s_2$, \dots, $s_\text{DIM-1}$,
            $s_\text{DIM+1}$, \dots $s_n$) of NORM2(X,DIM=DIM[,MASK=MASK])
            is equal to
               NORM2(X($s_1$, $s_2$, \dots, $s_\text{DIM-1}$, :,
                     $s_\text{DIM+1}$, \dots $s_n$)
                     [,MASK=MASK($s_1$, $s_2$, \dots, $s_\text{DIM-1}$, :,
                                 $s_\text{DIM+1}$, \dots $s_n$)])."

[457:39-31 16.9.153p7 Example] Replace the paragraph by

"Examples.

"Case(i): The value of NORM2([3.0,4.0]) is 5.0 (approximately).

                              [ 1.0 2.0 ]
"Case(ii): If X has the value [         ] then the value of
                              [ 3.0 4.0 ]

           NORM2(X,DIM=1] is [3.162,4,472] (approximately) and the value
           of NORM2(X,DIM=2) is [2.236,5.0] (approximately).

"Case(iii): If X has the value [ -3.0 -99.0 4.0 ] then the result of
            NORM2(X,MASK=(X > -99.0)) has the value 5.0 (approximately)."

[457:31+ 16.9.153p7+ Example] Insert a note:

"It is recommended that the processor use a method that does not cause
overflow or underflow during the calculation if the value of the result
does not underflow or overflow."

! One way that can be done is to compute Y = X / MAXVAL(ABS(X)) and
! calculate MAXVAL(ABS(X)) * NORM2(Y). A more efficient way is to
! proceed using the obvious method, and switch to the first method only
! if an overflow or underflow exception occurs. User programs cannot
! detect or handle exceptions.