**-011 [intro.defs] Term not used in the document

[jensmaurer]

terms defined in 3.2, 3.18, 3.29, 3.66

Please use this term in the document or remove it from Clause 3. ISO/IEC Directives, Part 2, 16.5.4: “Only terms which are used in the document shall be listed in the Terms and definitions clause.”

[jensmaurer]

Removing should be fine if actually unused.

[jensmaurer]

3.18 is "direct-non-list-initialization", which is used pervasively in the verb form "direct-non-list-initialized".

3.29 the term "iostream class" is used twice, but not in a way that would need a definition here. Removed.

3.66 "unspecified behavior" is used pervasively as "it is unspecified whether ..."

3.2 "arbitrary-positional stream" and 3.48 "repositional stream" are removed, because unused.

[jensmaurer]

We need a decision for

3.18 is "direct-non-list-initialization", which is used pervasively in the verb form "direct-non-list-initialized". 3.66 "unspecified behavior" is used pervasively as "it is unspecified whether ..."

[JohelEGP]

3.66 "unspecified behavior" is used pervasively as "it is unspecified whether ..."

The term should be "unspecified", or the use like "it is unspecified behavior whether ...". Remember this from the directives:

16.5.6 Definitions

The definition shall be written in such a form that it can replace the term in its context.

[tkoeppe]

Decision: delete the three; reword "direct-non-list-initialization", keep "unspecified behaviour".

[tkoeppe]

@jensmaurer, @zygoloid ISO really wants us to remove "unspecified behavior".

Could we maybe change it to just define "unspecified", and say:

<behavior> unspecified: correct and dependent on the implementation

[jensmaurer]

Suggestion: Let's rephrase some "unspecified" thing in the standard proper to use the phrasing "unspecified behavior". Then ISO/CS is happy, right?

[jensmaurer]

Suggestion in linked pull request.

[tkoeppe]

This is nice, thanks!