Fix modal verbs #4194

jensmaurer · 2020-09-13T19:06:47Z

Partially addresses cplusplus/nbballot#393

zygoloid

Thanks! I think we can be a bit bolder in using "can" rather than "could"; usually there doesn't seem to be any real difference in the result, and it seems like ISO is a lot happier about "can" than "could".

zygoloid · 2020-09-15T23:51:55Z

source/algorithms.tex

+Consequently, the algorithms cannot
 be called with explicitly-specified template argument lists.


Hm. "cannot portably" would be more accurate, but I don't know that we want to talk about portability. How about:

Suggested change

Consequently, the algorithms cannot

be called with explicitly-specified template argument lists.

Consequently, an implementation can reject

calls that specify an explicit template argument list.

zygoloid · 2020-09-15T23:57:14Z

source/algorithms.tex

@@ -9537,7 +9537,7 @@
 that \tcode{exclusive_scan} excludes the $i^\text{th}$ input element
 from the $i^\text{th}$ sum.
 If \tcode{binary_op} is not mathematically associative,
-the behavior of \tcode{exclusive_scan} may be nondeterministic.
+the behavior of \tcode{exclusive_scan} is nondeterministic.


Suggested change

the behavior of \tcode{exclusive_scan} is nondeterministic.

the behavior of \tcode{exclusive_scan} might be nondeterministic.

I think whether it actually is deterministic or not depends on the implementation. We certainly don't want to imply that an implementation is required to have a source of true nondeterminism. (Use "can be" if you prefer.)

zygoloid · 2020-09-15T23:58:39Z

source/algorithms.tex

@@ -9664,7 +9664,7 @@
 that \tcode{inclusive_scan} includes the $i^\text{th}$ input element
 in the $i^\text{th}$ sum.
 If \tcode{binary_op} is not mathematically associative,
-the behavior of \tcode{inclusive_scan} may be nondeterministic.
+the behavior of \tcode{inclusive_scan} is nondeterministic.


Suggested change

the behavior of \tcode{inclusive_scan} is nondeterministic.

the behavior of \tcode{inclusive_scan} might be nondeterministic.

zygoloid · 2020-09-15T23:58:49Z

source/algorithms.tex

@@ -9739,7 +9739,7 @@
 \tcode{transform_inclusive_scan} is that \tcode{transform_exclusive_scan}
 excludes the $i^\text{th}$ input element from the $i^\text{th}$ sum.
 If \tcode{binary_op} is not mathematically associative,
-the behavior of \tcode{transform_exclusive_scan} may be nondeterministic.
+the behavior of \tcode{transform_exclusive_scan} is nondeterministic.


Suggested change

the behavior of \tcode{transform_exclusive_scan} is nondeterministic.

the behavior of \tcode{transform_exclusive_scan} might be nondeterministic.

zygoloid · 2020-09-15T23:59:00Z

source/algorithms.tex

@@ -9843,7 +9843,7 @@
 \tcode{transform_inclusive_scan} is that \tcode{transform_inclusive_scan}
 includes the $i^\text{th}$ input element in the $i^\text{th}$ sum.
 If \tcode{binary_op} is not mathematically associative,
-the behavior of \tcode{transform_inclusive_scan} may be nondeterministic.
+the behavior of \tcode{transform_inclusive_scan} is nondeterministic.


Suggested change

the behavior of \tcode{transform_inclusive_scan} is nondeterministic.

the behavior of \tcode{transform_inclusive_scan} might be nondeterministic.

zygoloid · 2020-09-16T01:08:31Z

source/templates.tex

@@ -3696,7 +3696,7 @@
 the relationship between the template parameters and the rest of the
 signature.
 \begin{note}
-Two distinct function templates may have identical function return types and
+Two distinct function templates could have identical function return types and


Suggested change

Two distinct function templates could have identical function return types and

Two distinct function templates can have identical function return types and

zygoloid · 2020-09-16T01:09:16Z

source/templates.tex

@@ -7717,7 +7717,7 @@

 \pnum
 \begin{note}
-Type deduction may fail for the following reasons:
+Type deduction could fail for the following reasons:


Suggested change

Type deduction could fail for the following reasons:

Type deduction can fail for the following reasons:

zygoloid · 2020-09-16T01:10:08Z

source/time.tex

@@ -1058,7 +1058,7 @@
 \pnum
 The \tcode{duration} template uses the \tcode{duration_values} trait to
 construct special values of the duration's representation (\tcode{Rep}). This is
-done because the representation might be a class type with behavior which
+done because the representation can be a class type with behavior which


Suggested change

done because the representation can be a class type with behavior which

done because the representation can be a class type with behavior that

(en passant)

zygoloid · 2020-09-16T01:10:38Z

source/time.tex

@@ -1140,7 +1140,7 @@
 \tcode{duration} arguments will convert to it without requiring a division
 operation. The representation of this type is intended to be able to hold any
 value resulting from this conversion with no truncation error, although
-floating-point durations may have round-off errors.
+floating-point durations could have round-off errors.


Suggested change

floating-point durations could have round-off errors.

floating-point durations can have round-off errors.

source/lex.tex

jensmaurer · 2020-09-16T09:17:12Z

Rebased, fixed + force-pushed.

source/atomics.tex

zygoloid · 2020-09-17T22:53:35Z

source/basic.tex

+  \recommended
+    Implementations should ensure that
+    repeated and particularly inopportune interference
+    from other threads, e.g. by repeatedly stealing a cache line


Suggested change

from other threads, e.g. by repeatedly stealing a cache line

from other threads, for example by repeatedly stealing a cache line

zygoloid · 2020-09-17T23:01:02Z

source/basic.tex

+  \recommended
+    Implementations should ensure that
+    repeated and particularly inopportune interference
+    from other threads, e.g. by repeatedly stealing a cache line
    for unrelated purposes


I don't think this is a new, independent "should", but a restatement of the prior normative "should". Maybe something like this would be an improvement:

[Note: Under this condition, interference from other threads cannot indefinitely delay progress, even if such interference is repeated and particularly inopportune (for example, by repeatedly stealing a cache line between load-locked and store-conditional instructions), and such anomalies can therefore be safely ignored by programmers. -- end note]

"Under this condition" sounded not good. Rephrased by putting programmers first; please check again.

zygoloid · 2020-09-17T23:01:57Z

source/basic.tex

-    may prevent forward progress,
-    e.g.,
-    by repeatedly stealing a cache line
+  \recommended


I don't think we want a \recommended here. I think the text that follows is just an elaboration of the should on the previous two lines.

zygoloid · 2020-09-17T23:16:10Z

source/containers.tex

 \tcode{vector}
-is the type of sequence container that should be used by default.
+is the sequence container that can be used by default.


I don't think "can" is appropriate here; this really is advice for the programmer. Maybe we can avoid normative encouragement with factual statements like: "The sequence container vector is appropriate in most circumstances. Use of array may be appropriate when the container has a fixed size [...]"

Removed "can", but otherwise left the structure intact. Note the funny part right afterwards, for which we should leave the context intact.

jensmaurer · 2020-09-18T20:13:10Z

Fixed, rebased, and force-pushed.

jensmaurer · 2020-09-18T20:40:20Z

Added a commit that removes should and may from footnotes (simple "grep").

zygoloid · 2020-09-18T23:16:59Z

Hi Jens, there are quite a lot of unaddressed comments still. The github UI is pretty bad and sometimes hides comments in the "Conversation" view, but you can manually unhide them, and most of them (those that still match a line in the current set of changes) will show up in the "Files changed" view.

jensmaurer · 2020-09-19T20:55:37Z

@zygoloid , I found the rest of the comments and hopefully addressed them all.

jensmaurer · 2020-09-20T11:27:13Z

Some more fixes for footnotes.

source/atomics.tex

source/future.tex

source/iostreams.tex

source/threads.tex

source/utilities.tex

jensmaurer · 2020-09-21T22:04:29Z

All fixed + pushed.

Partially addresses ISO/CS 017 (C++20 DIS)

and prefix them with "Recommended practice" Partially addresses ISO/CS 017 (C++20 DIS)

Partially addresses ISO/CS 017 (C++20 DIS)

zygoloid added this to the C++20 milestone Sep 15, 2020

zygoloid added the ballot-comment Response to an NB or ISO comment on a ballot label Sep 15, 2020

zygoloid reviewed Sep 16, 2020

View reviewed changes

zygoloid added the needs rebase The pull request needs a git rebase to resolve merge conflicts. label Sep 16, 2020

jensmaurer force-pushed the cs017 branch from 9ba035b to 6c5a4c4 Compare September 16, 2020 09:17

jensmaurer removed the needs rebase The pull request needs a git rebase to resolve merge conflicts. label Sep 16, 2020

jensmaurer mentioned this pull request Sep 16, 2020

CS017 Strictly adhere to permitted modal verb forms cplusplus/nbballot#393

Closed

jensmaurer force-pushed the cs017 branch 2 times, most recently from 6a86bab to 1154577 Compare September 17, 2020 15:27

zygoloid reviewed Sep 18, 2020

View reviewed changes

zygoloid added the changes requested Changes to the wording or approach have been requested and not yet applied. label Sep 18, 2020

jensmaurer force-pushed the cs017 branch 2 times, most recently from 72c2bfa to 90f630a Compare September 18, 2020 20:11

jensmaurer removed the changes requested Changes to the wording or approach have been requested and not yet applied. label Sep 18, 2020

jensmaurer force-pushed the cs017 branch from 515e84f to 7c2d9b2 Compare September 18, 2020 20:42

zygoloid added the changes requested Changes to the wording or approach have been requested and not yet applied. label Sep 18, 2020

jensmaurer force-pushed the cs017 branch from 7c2d9b2 to 5cbf68a Compare September 19, 2020 20:55

jensmaurer removed the changes requested Changes to the wording or approach have been requested and not yet applied. label Sep 19, 2020

jensmaurer force-pushed the cs017 branch from 05911ab to 1a93d4a Compare September 21, 2020 21:20

zygoloid reviewed Sep 21, 2020

View reviewed changes

jensmaurer added 3 commits September 22, 2020 00:23

[std] Replace 'might' in normative context.

73c418e

Partially addresses ISO/CS 017 (C++20 DIS)

[std] Replace 'may' in notes.

9e2dd1a

Partially addresses ISO/CS 017 (C++20 DIS)

[std] Replace 'must'.

4718d20

Partially addresses ISO/CS 017 (C++20 DIS)

jensmaurer added 4 commits September 22, 2020 00:34

[std] Move implementation recommendations to outside notes

2186553

and prefix them with "Recommended practice" Partially addresses ISO/CS 017 (C++20 DIS)

[std] Rephrase notes giving advice to the programmer.

527a5a6

Partially addresses ISO/CS 017 (C++20 DIS)

[check] Flag 'shall', 'may', or 'should' inside notes.

7b622b9

[std] Remove 'should' and 'may' from footnotes.

02a3bf0

Partially addresses ISO/CS 017 (C++20 DIS)

jensmaurer force-pushed the cs017 branch from a605e4d to 02a3bf0 Compare September 21, 2020 22:34

zygoloid merged commit 550bcf5 into cplusplus:c++20 Sep 21, 2020

jensmaurer deleted the cs017 branch September 21, 2020 22:37

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Fix modal verbs #4194

Fix modal verbs #4194

jensmaurer commented Sep 13, 2020 •

edited

zygoloid left a comment

zygoloid Sep 15, 2020

jensmaurer Sep 16, 2020

zygoloid Sep 15, 2020

jensmaurer Sep 16, 2020

zygoloid Sep 15, 2020

jensmaurer Sep 16, 2020

zygoloid Sep 15, 2020

jensmaurer Sep 16, 2020

zygoloid Sep 15, 2020

jensmaurer Sep 16, 2020

zygoloid Sep 16, 2020

jensmaurer Sep 16, 2020

zygoloid Sep 16, 2020

jensmaurer Sep 16, 2020

zygoloid Sep 16, 2020

jensmaurer Sep 16, 2020

zygoloid Sep 16, 2020

jensmaurer Sep 16, 2020

jensmaurer commented Sep 16, 2020

zygoloid Sep 17, 2020

jensmaurer Sep 18, 2020

zygoloid Sep 17, 2020

jensmaurer Sep 18, 2020 •

edited

zygoloid Sep 17, 2020

jensmaurer Sep 18, 2020

zygoloid Sep 17, 2020

jensmaurer Sep 18, 2020

jensmaurer commented Sep 18, 2020

jensmaurer commented Sep 18, 2020

zygoloid commented Sep 18, 2020

jensmaurer commented Sep 19, 2020 •

edited

jensmaurer commented Sep 20, 2020

jensmaurer commented Sep 21, 2020

		Consequently, the algorithms cannot
		be called with explicitly-specified template argument lists.

	the behavior of \tcode{exclusive_scan} is nondeterministic.
	the behavior of \tcode{exclusive_scan} might be nondeterministic.

	the behavior of \tcode{inclusive_scan} is nondeterministic.
	the behavior of \tcode{inclusive_scan} might be nondeterministic.

	the behavior of \tcode{transform_exclusive_scan} is nondeterministic.
	the behavior of \tcode{transform_exclusive_scan} might be nondeterministic.

	the behavior of \tcode{transform_inclusive_scan} is nondeterministic.
	the behavior of \tcode{transform_inclusive_scan} might be nondeterministic.

	Two distinct function templates could have identical function return types and
	Two distinct function templates can have identical function return types and

	Type deduction could fail for the following reasons:
	Type deduction can fail for the following reasons:

	done because the representation can be a class type with behavior which
	done because the representation can be a class type with behavior that

	floating-point durations could have round-off errors.
	floating-point durations can have round-off errors.

	from other threads, e.g. by repeatedly stealing a cache line
	from other threads, for example by repeatedly stealing a cache line

Fix modal verbs #4194

Fix modal verbs #4194

Conversation

jensmaurer commented Sep 13, 2020 • edited

zygoloid left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jensmaurer commented Sep 16, 2020

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jensmaurer Sep 18, 2020 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jensmaurer commented Sep 18, 2020

jensmaurer commented Sep 18, 2020

zygoloid commented Sep 18, 2020

jensmaurer commented Sep 19, 2020 • edited

jensmaurer commented Sep 20, 2020

jensmaurer commented Sep 21, 2020

jensmaurer commented Sep 13, 2020 •

edited

jensmaurer Sep 18, 2020 •

edited

jensmaurer commented Sep 19, 2020 •

edited