To: J3                                                     J3/23-153
From: Malcolm Cohen
Subject: Editor's report on FDIS production.
Date: 2023-April-12


1. Introduction

This is a report of the changes made between the DIS and the FDIS.

Paragraph numbers are *not* added to the FDIS, to reduce friction with ISO.
(They are planned to be present in the Interpretation document.)


2. List of papers in numerical order

23-100r2    23-137
23-128      23-138r2
23-132      23-139
23-133      23-140
23-134      23-142
23-135      23-144r1
23-136      23-149


3. Papers applied in order

23-135.

Produced 23-007 with line numbers and paragraph numbers.

Done.

23-136 partial.

Just the "pi upright" changes.

Done.

23-132.

Done.

23-133.

Done.

23-134.

COMMENT: This makes 9.7.1.2p4 even more wall-of-texty.
         It's not obvious how to improve it though.

Done.

23-137.


COMMENT: "a reference to a function that returns a data pointer" actually
         is a (noncoindexed) variable, so that part of the requirement is
         redundant. Not that more than one compiler has implemented that
         F2008 feature correctly!

EXTRA EDIT: Hyperlinked "intrinsic assignment" throughout clause 16
            (about half were hyperlinked, the others not).

EXTRA EDIT [348:9] 16.6 Collective subroutines, p8,
    "assigned an explanatory message by intrinsic assignment."
    ->"assigned an explanatory message, as if by intrinsic assignment."

Done.

23-128.

Done.

23-136 (apart from upright pi).

EXTRA EDIT [154:25+1]: Caption is ungrammatical: disagreement in number,
    viz "Type of operands and results...", there is more than one type
    involved, so it should be "Types of operands and ...". Fixed.

COMMENT: Increased longtable caption width from 12.7cm to 14cm so that
         Table 10.6 caption still fitted. (An emdash and an extra space are
         a lot wider than a colon.)

COMMENT: Table 18.3 ditto.

Done.

23-100r2.

Done.

23-139.

DIFFERENT throughout: properly capitalised the sentence.

DIFFERENT: [6:3,4+1] inserted comma before "of a...".

DIFFERENT: [14:15] deleted "and".

Done.

23-140.

Edit:
[Clause 3 throughout] Change intra-clause hyperlinks to have the term in
                      italics, followed by its term number in parentheses.
                      Both parts should be actively hyperlinked.

COMMENT: All of these were changed to be non-indexing.

DIFFERENT: Instead of doing this I "unlinked" unimportant terms, in:
    array - "type parameters"
    explicit-shape array - "bounds"
    use association - "scoping unit"
    declaration - in the note, "data object" and "object"
    default initialization - "pointer association"
    definition "derived type" and "procedure" (in the qualification part)
    disassociated - everything "pointer association", "target", "undefined"
    finalization - "final subroutines"
    argument keyword - "actual argument"
                       because it's really "actual argument list"
    pointer assignment - "intrinsic" (before "assignment")
    module procedure - "intrinsic"
    data object reference - "data object designator"
    sequence structure - "scalar"
    structure - "scalar"
    atomic subroutine - "intrinsic"
    collective subroutine - "intrinsic"
    transformational function - "intrinsic" twice
    derived type - "intrinsic"
    parent type - "extended type" (in the NOTE)
    undefined - "data object" and "pointer" (in the qualifications)
    undefined (pointer) - "disassociated"

DIFFERENT: "pointer associated" is not itself a term, so that remains a
           hyperlink to "pointer association" (without a cross-reference).
           Similarly "assumed" (assumed type parameter)
           and "deferred" (deferred type parameter)
           "pointer association status" is not itself a term, so I unlinked
           it from "pointer association".

DIFFERENT: I had to move the cross-ref at the end of the entry into a NOTE
           in the following cases, because it interfered with the cross-ref
           insisted on by ISO; NOTE says "XYZ is/are described in A.B.C."
    assumed-shape array
    assumed-size array
    deferred-shape array
    argument association
    inheritance association
    name association
    automatic data object
    binding label
    component order
    external unit
    file storage unit
    image control statement
    interface block
    interoperable
    internal file
    internal unit (actually I put both under internal file)
    segment
    storage units
    submodule
    type compatible (and changed the ref from 7.3.2 to 7.3.3)

EXTRA EDIT: Rewrote "associating entity"
            (I'm sure he merely overlooked the qualification last time).

EXTRA EDIT: Deleted cross-reference(s) at the end of an entry when it can
            be found by following the ISO-mandated references for the
            immediately preceding terms, in
    binding
    bound/array bound
    coarray
    cobound
    codimension
    component keyword
    procedure reference
    storage sequence
    structure

EXTRA EDIT: In "automatic data object",
            changed syntax term <specification-expr>
            to defined term "specification expression" (with the ref).
            That's because it was wrong before - the expression that makes
            it automatic is not always accessed through the syntax term
            (e.g. constraint C805 requires type-param-value in a type
            declaration statement to be a specification expression).

EXTRA EDIT: In "component order", move the conditional clause before the
            one it is acting on, (otherwise the ref interferes).

EXTRA EDIT: rewrote definition of "elemental procedure",
            as the existing text was incomplete i.e. wrong.

EXTRA EDIT: rewrote definition of "generic identifier" as the existing text
            was wrong.

EXTRA EDIT: for "host instance", I just deleted "of an internal procedure"
            because I did not want to use the term reference syntax, and
            that bit is explained in the following NOTE anyway.

EXTRA EDIT: in "dummy procedure" reverse ordering so the good ref still
            appears properly.

EXTRA EDIT: in "rank", rewrite to make better sense with the mandatory ref.

EXTRA EDIT: in "subobject",move the commas around to improve the grammar.

EXTRA EDIT: in "current team",
            move the cross-reference to "CHANGE TEAM construct"
            to immediately follow the first textual "CHANGE TEAM construct"
            as I think that is clearer.
            Similarly in "parent team" for "FORM TEAM statement".

EXTRA EDIT: Changed 7.3.2 title
                from "Type specifiers and type compatibility"
                to "Type specifiers",
            because 7.3.3 is "Type compatibility", and the words compatible
            or compatibility do not appear in 7.3.2 at all.

Done.

23-142.

Done.

23-144r1.

Done.

23-138r2.

The replacement foreword was discovered, and our version updated.

Done.

23-149.

EXTRA EDIT: 7.1.4 Constants, p3, index "array constructor".

Done.

Private Communication from John Reid.

EXTRA EDIT: In ATAN2PI, change "degrees" to "half-revolutions".

Done.

Changes for the FDIS.

N2213.

Now to go through the whole document and find widows/orphans...

EXTRA EDIT: Table 16.4 cannot be split across a page, as the caption is
            destroyed viz "(cont.)" overwrites the caption.
            Kludged the column widths to avoid that.

COMMENT: Table 18.3 was similar, but I managed to squeeze it onto the page.

===END===