The interface to the IETF document process is exactly what I use xml2rfc for.
If I am producing a -bis document, I pull the .xml from wherever I can find it and convert it to word. Then I convert it back to upload it.
But since we are discussing all this. How about we specify the rest of the interface. I would really like it if I could make the document submission from RFCtool.
So normally, I type
rfctool %DocSource%\hallambaker-mesh-5-protocol.docx /auto /cache=bib.xml
It would be nice to be able to submit a changed document with something like
rfctool %DocSource%\hallambaker-mesh-5-protocol.docx /auto /cache=bib.xml /submit
There is a small corner case which is that if you have multiple documents and you are submitting updates to three out of five, you want the new docs to reference the new versions of the updates documents but the old versions of the ones that are not being submitted.
So it is probably something a bit more git-like and there is a /stage operation to mark the draft as one of the ones to be updated and a /submit operation to ship them out.
[I would also prefer to be able to authenticate the submission by signing it with a digital signature rather than the email callback scheme. Which might sound like feature creep but could actually simplify providing the capability as it would allow a stateless submission. The XML would have to be expanded to specify the signature key(s) allowed.]
On Thu, Jul 11, 2019 at 1:32 PM Keith Moore <moore@xxxxxxxxxxxxxxxxxxxx> wrote:
Sent from my iPhone
> On Jul 11, 2019, at 1:24 PM, Nico Williams <nico@xxxxxxxxxxxxxxxx> wrote:
>
>> On Thu, Jul 11, 2019 at 09:41:09AM -0400, Phillip Hallam-Baker wrote:
>> But the bigger point here surely, is that this thread is now an editor war.
>
> Sigh. That was not my intention, but the opposite. My point was/is
> that the choice of $EDITOR shouldn't matter, and that our choice of XML
> as the canonical or submission format should not -and DO NOT- preclude
> $EDITORs that hide the XML from the user.
Actually the one thing that xml seems like it might be a decent choice for, is a stable back end interface that could accept input from a variety of editors and text processors and allow RFCs and I-Ds to be handled by a common toolset. Except that it’s not stable.
(It’s not a particularly good interface for that purpose, but it doesn’t have to be - the main requirements are that it be adequate and stable.)
Keith
