To: J3                                                     J3/17-177r1
From: Dan Nagle
Subject: Reply to Public Comments
Date: 2017 June 27


Introduction
************

This paper provides WG5 responses to some public comments of 17-007r1.
Comments 12, 13, 14, 15, 23, 24, 25, 26, 30, 31, 32, and 44 are here.

**

Comment GB012
The argument of LOG10 needs to be a real. The result of RADIX is integer.

Response:  We agree.  The suggested edits are recommended in this paper.

**

Comment GB013
y has not been declared.

Response:  We agree.  The suggested edits are recommended in this paper.

**

Comment GB014
lb1, ub1, lb2 are not declared to be integer

Response:  We agree.  Edits are supplied in this paper.

**

Comment GB015
IF (ANY(a(:,j))==0) EXIT" is not correct

Response:  We agree.  The suggested edits are recommended in this paper.

**

Comment GB023
The hexadecimal indicator 0X should appear before
the first digit and it does in NOTE 13.14.
See also 13.7.2.3.2 para 7.

Response:  We agree.  The suggested edits are recommended in this paper.

**

Comment GB024
The choice of binary exponent is processor dependent.

Response:  We agree.  The suggested edits are recommended in this paper.

**

Comment GB025
The final value in NOTE 13.14 does not match the first value
on the line.

Response:  We agree.  The suggested edits are recommended in this paper.

**

Comment GB026
To illustrate that the leading hex digit need not be 1,
add another example to NOTE 13.14.

Response:  We agree.  The suggested edits are recommended in this paper.

**

Comment GB030
The actual arguments corresponding to the arguments MOLD and ROUND
of the elemental intrinsic function OUT_OF_RANGE are required
to be scalar, just as for dummy arguments KIND of other elementals,
but this is not said in 16.2.1.

Response: We disagree; the arguments are specified to be scalar.
However, to clarify the issue identified, we will add a note
immediately following 16.9.146.

**

Comment GB031
The result of OUT_OF_RANGE is incorrectly limited to being scalar.

Response:  We agree.  The suggested edits are recommended in this paper.

Note that the elemental function result is scalar.
The effect of referencing the elemental function has the shape
of the possibly array actual argument.  However, other examples
of intrinsic elemental functions where the result is not the type and kind
of a possibly array actual argument do not specify scalar.  For example,
see LEN_TRIM.  The same wording should be used here.

**

Comment GB032
The list of arguments does not include KIND.

Response: We agree.  The suggested edits are recommended in this paper.

**

Comment GB044
The choice of binary exponent in EX output editing is processor dependent.

Response: We agree.  The suggested edits are recommended in this paper.


Edits to 17-007r1
*****************

{in Note 10.34 [168-169] just before 10.2 Assignment}

In Note 10.34, change the line:
"CEILING (number_of_decimal_digits / LOG10 (RADIX (0.0)))"
to
"CEILING (number_of_decimal_digits / LOG10 ( REAL (RADIX (0.0))))"

{show the correct return type of radix()}
{GB012}

{11.1.10.4 Examples of the SELECT RANK construct}

[202:23] change (y) to (x) twice, so
"Print *, 'I did not expect rank', RANK(y), 'shape', SHAPE(y)"
changes to
"Print *, 'I did not expect rank', RANK(x), 'shape', SHAPE(x)"

{make the example correct code}
{GB013}

{11.1.10.4 Examples of SELECT RANK construct}

[202:46+] show declarations of the integers add a line
"INTEGER, INTENT (IN) :: lb1, ub1, lb2"

{put the declarations in the example}
{GB014}

{11.1.10.4 Examples of the SELECT RANK construct}

[203:6] change
"IF (ANY(a(:,j))==0) EXIT"
to
"IF (ANY(a(:,j)==0)) EXIT"

{correct the example}
{GB015}

{13.7.2.3.6 EX editing}

[278:15] change
"[+-]x_0.x_1x_2...exp"
to
"[+-]0Xx_0.x_1x_2...exp"

{show the correct leading 0X}
{GB023}

{13.7.2.3.6 EX editing}

[278:23] after the sentence "...zero the form is P +-z_1...z_e."
add to the bullet
"The choice of binary exponent is processor dependent.
If the most significant binary digits of the internal value
are b_0b_1b_2..., the binary exponent might make the value
of x_0 be that of b_0, b_0b_1, b_0b_1b_2, or b_0b_1b_2b_3."

{explain the implications of processor choice of radix}
{GB024}

{13.7.2.3.6 EX editing}

[278:24+] in Note 13.14, on the final line,
change "0X1.00003P+20" to "0X1.00004P+20"

{show correct hex value}
{GB025}

{13.7.2.3.6 EX editing}

[278:24+] in Note 13.14, add a line at the end of the note

"2.375  EX0.1   0X2.6P+0"

{show an example where the leading digit is > 1}
{GB026}

{16.9.55 COSHAPE (COARRAY [, KIND])}

[373:25+] add a line
"KIND (optional) shall be a scalar integer constant expression."

{describe all dummy arguments}
{GB032}

{16.9.146 OUT_OF_RANGE (X, MOLD [, ROUND])

[416:17] delete "scalar"

{match wording of len_trim}
{GB031}

{add a note saying why mold and round are scalar}

[416:31+]
"Note 16.23+
MOLD is required to be a scalar because the only information
taken from it is its type and kind.  Allowing an array MOLD would
require that it be conformable with X.  ROUND is scalar because
allowing an array rounding mode would have severe performance
difficulties on many modern processors."

{explain the scalar-ness of mold and round}
{GB030}

{A.2 Processor Dependencies}

{Note that exponent choice is PD}

[538:39+] add a sentence

"the choice of binary exponent in EX output editing (13.7.2.3.6)."

{note the PD of radix}
{GB044}

===EOF===