To: J3                                                     J3/23-183
From: Malcolm Cohen
Subject: Notes about 23-007r1
Date: 2023-June-13

This is a guide to the changes (mostly but not entirely in clause 3)
which are in 23-007r1.

DELETED term "nonstandard intrinsic".
Readers will understand this without a term.

A few differences where I joined lines.

All references in Clause 3 changed to "non-indexing" ones.
Note: The definitions themselves are still referenced.

Nearly all references on lines made into "Note N to entry".
I deleted some I thought were useless/unimportant,
or already appeared in a "parent" term.
If there was already a Note there, I usually did not join them
i.e. the second one became "Note 2 to entry".

Totally rewrote the multitudinous "characteristics" into a single
one with a big note.

Occasionally the new Note just says "See ...", but mostly I tried
to make it a real sentence not a command. Resorted to "See..." when
no obviously-good real sentence sprang to mind. Mostly used the verb
"described", but used "defined" a few times when I thought it right.

Sometimes left the old text there commented-out (lines beginning with "%").

Turned BNF image-selector in the cosubscript definition into text,
with a Note saying what its syntax is.

DELETED the term "NaN", instead making it a synonym of "IEEE NaN",
which term ISO did not complain about.

After "active image", just deleted the ref to "Image execution states"
as surely people can manage to read a few lines higher.

Similarly, Note for "data pointer",
but not the following "procedure pointer".

SPONTANEOUS EDIT: Added ".." to the list in "lexical token".

DELETED the term "standard-conforming program", as it was only linked
within 4.2 Conformance, which repeated the definition anyway.

Simply deleted the ref on "character storage unit" et al, as it is in
the "storage unit" parent definition.

Hyperlinked "EXTENDS" in "parent type" definition.

INDEXED ".. assumed-rank specifier".

FIXED mistake in SHAPE, which prohibited assumed-rank despite saying
what the result was for assumed-rank.

NOTE There are a lot of "may be scalar or an array" in the inquiry
functions, which is not technically wrong (so I did not fix them) but
are definitely misleading. Just deleting the permission is the right
thing to do in all of these.

FIXED mistake in IEEE_MIN_NUM, which conflicted with the IEEE std.

FIXED mistake in IEEE_RINT, which gave the wrong name for the IEEE
function.

===END===