Re: [xml2rfc] Insufficiency of txt format

[Paul Kyzivat <[e-mail]>]

On 1/31/21 9:35 PM, John Levine wrote:
> In article <[e-mail]>,
> Paul Kyzivat  <[e-mail]> wrote:
>> Of more concern, AFAIK the only reasonable diff tools work on the txt
>> version. IIRC there was once talk of creating a diff tool that worked on
>> the xml, but I haven't heard of any such thing materialized.
> 
> Jim Schaad was working on it, but the stuff he wrote wasn't close to done.
> 
>> I think
>> that would present great challenges, especially if the xml is
>> aesthetically revised while preserving the look. And I would still want
>> to see the diff results in side-by-side form. Also that won't work when
>> one of the versions being diffed predates v3 xml.
> 
> I agree there's no plausible way to do XML diffs on documents using
> different XML vocabularies, but eventually v2 will die out. Our
> tooling is a problem and we underestimated what we'd need beyond
> xml2rfc to create a productive v3 xml environment.

I'll be dead before v2 xml is.

I worked on rfc4566bis for a long time, and it was just published as 
rfc8866. Is is very informative to look at the diff between the two.

And the txt diff will remain an important tool as long as there are 
pre-v3 documents that get bis versions.

>> The only way I have seen to reasonably manage differences in documents
>> with fancier formatting is with change history, such as provided by
>> MSWord. We clearly don't have that with v3 xml.
> 
> Well, we do if we store them in github but that's its own can of worms.

That will only show the changes in the xml. That isn't generally what I 
want to see when reviewing a new revision of a document. The 
side-by-side txt diff usually gives me exactly what I want to see when 
commenting on a new version. It would be nice if I could get comparable 
functionality between html or pdf renderings of the document, but I 
can't imagine how that would work. (Hopefully this is just a statement 
about my lack of imagination.)

	Thanks,
	Paul