Re: [xml2rfc-dev] https://tools.ietf.org/html/draft-levkowetz-xml2rfc-v3-implementation-notes-09#section-3.1.20

[Julian Reschke <[e-mail]>]

On 29.10.2019 19:13, Heather Flanagan wrote:
>
>
>> On Oct 29, 2019, at 11:09 AM, Julian Reschke <[e-mail]> wrote:
>>
>> ...see also
>> <https://github.com/rfc-format/draft-iab-xml2rfc-v3-bis/pull/107>, which
>> I believe is misguided.
>>
>>
>>> 3.1.20.  In Section 2.46.2, "numbered" Attribute
>>>
>>>    The text indicates that only top-level sections may have
>>>    numbered="false", and that a section with numbered="false" may not
>>>    have a child section with numbered="true".  But that leaves no value
>>>    that is valid for child sections of an unnumbered section: They
>>>    cannot have numbered="false", since they are not top-level sections,
>>>    and they cannot have numbered="true", since the parent has
>>>    numbered="false".
>>
>> That is correct, but very easy to fix by saying that numbering is
>> inherited from the parent section.
>>
>>>    Additionally, the prohibition against child sections having
>>>    numbered="false" removes the option of truncating the ToC listing for
>>>    some child sections; without providing a good explanation for this
>>>    limitation, it seems arbitrary and counter-intuitive to disallow this
>>>    feature.
>>
>> That doesn't make any sense - the presence in the ToC does not depend on
>> numbering at all. To control the ToC, we already have
>> <https://greenbytes.de/tech/webdav/rfc7991.html#element.section.attribute.toc>.
>>
>>>    Proposal:  Permit sections which are not top-level sections to have
>>>       numbered="false".
>>>
>>>    Implementation:  In The current version of xml2rfc, child sections
>>>       may have numbered="false".
>>
>> Limiting this to top-level sections was very intentional; the reasons
>> given above are IMHO not sufficient at all for this change.
>>
>> Heather, by any means, please undo this change for now.
>>
>>
>
> I don’t think having numbered child sections in an otherwise unnumbered section (especially if they are inheriting a phantom number from the unnumbered section” is a good idea at all. Do you have better language that I can use to state that is prohibited?

Just say that descendant sections of unnumbered sections are unnumbered
by definition (and that it would be an error to set numbered back to true).

Asking authors to repeat themselves when the intent is totally clear
just doesn't make any sense.

FWIW, I'm also opposed to expanding control to subsections for now (and
I see that you didn't adopt that part of the proposal yet). Before we do
that, we should have a clear use case.

Best regards, Julian