[rfc-i] Pagination requirements

Joe Touch touch at isi.edu
Fri May 25 08:16:37 PDT 2012

On May 25, 2012, at 8:11 AM, Julian Reschke wrote:

> On 2012-05-25 17:04, Joe Touch wrote:
>>> It would be a pity if we lost containment.
>> Why?
>> I think we need to be very relaxed in our requirements for markup. I would expect - at most:
>> - front matter (title, authors, date, RFC number)
>> - section tags (as jump targets)
>> - figure/table tags
>> - reference tags
>> - internal references to the above should point to those tags
>> I don't see a reason for *requiring* any other markings - including, and especially, "containment" markings.
>> If anyone has a rationale for *needing* that, please explain (where I don't think "markup purity" is sufficient, BTW).
> Containment is not only simpler to process, it also ensures on the source level that the document structure makes sense.

It's also not supported by Word, AFAICT, and at least partly responsible for why WYSIWYG editors so easily corrupt 2629-XML files, in my experience.

> I recommend you google for "HTML outline algorithm" for attempts to extract a usable section tree from an HTML document.

This is interesting academically, but if this isn't what is mostly used (and hasn't been for a long time, AFAICT), why *introduce* it as a format requirement?

We should *minimize* the requirements, to reduce risk for the future.

I agree it would be nice to retain the entire doc structure for future authors, but other than authorship - which can be VERY easily re-formatted in a modern editor - there doesn't seem to be necessary utility to containment tagging.


More information about the rfc-interest mailing list