Re: [xml2rfc-dev] RFC 7991 issue #37: Schema Issue, RFC 7991, In Section 2.12, <br>

[Henrik Levkowetz <[e-mail]>]

Hi Julian,

On 2018-10-03 14:19, Julian Reschke wrote:
> On 10/3/2018 2:13 PM, Henrik Levkowetz wrote:
>> ...
>> Yes, I've run that through the current text processor, and I particularly
>> looked at it when working on table rendering.
>> 
>> Where is it you need <br> to make this come out right?
> 
> In the titles, at least if we want to reproduce what the RFC has.

Thank you for that.  It provides much better understanding of the case
which prompted the introduction of <br>.

Now, the new v3 feature which cost absolutely most extra work to
implement, by far, was the addition of table rowspan capability.

If it really is imperative to break a column title in one particular
place (and I agree it may be desirable) then why can't it be handled
by using rowspan for the other header cells, and two cells for the
particular column title that needs to be broken in a controlled manner?

And second, why is this a concern in a column header, and not, for instance
in the document title?

This is the result of a too long document title today (an actual example
as rendered by the v3 text renderer):

---
Network Working Group                                       H. Levkowetz
Internet-Draft                                              Elf Tools AB
Intended status: Informational                            3 October 2018
Expires: 6 April 2019


       Implementation notes for RFC7991, "The 'xml2rfc' Version 3
                              Vocabulary"
           draft-levkowetz-xml2rfc-v3-implementation-notes-04
----

This certainly could have been made better by use of <br>.

And I'm sure that I'm not able to predict all other cases where <br> might
be useful, so if we believe it would be useful in some contexts, we should
not limit it to only one narrow context which we've discovered so far.

> 
>> So the conclusion would be to eliminate <br>, then ...
>> 
>>>> Please note that my suggestion was to either remove <br> altogether, or
>>>> permit it in inline context consistently.  I don't care that much about
>>>> which choice we make, but I still haven't seen any concrete example that
>>>> shows why it makes sense to disallow it in one context, and not the other.
>>>
>>> My concern is that when people can't get the table output they want,
>>> they'll fall back to artwork, which isn't helpful. See
>>> <https://greenbytes.de/tech/webdav/rfc7541.html#huffman.code>.
>> 
>> Understood.  I still don't see the problem with permitting <br> generally --
>> the argument about going to artwork has some validity also for body text
>> if people can't get the output they want.
> 
> That is true, but I'm prepared to argue that if they want to enforce a 
> line break in running text, they are doing something wrong.

But can we be so sure of that, that it's right to enforce the current
limitation?  Had you thought of the case of a document title, above?

Might there not be other cases?  Maybe it would be better to permit it,
and maybe (at least in some cases) issue a warning?


Best regards,

	Henrik