On Jul 10, 2019, at 18:31, Nico Williams <nico@xxxxxxxxxxxxxxxx> wrote: > >> don’t know of one that would work. Using markdown, or Latex, or >> whatever, just adds one more layer of translation you have to deal >> with later. I occasionally try to use markdown, but keep reverting >> back to XML because it’s just one less step. > > I agree as to markdown. Markdown we might never be able to > programmatically extract metadata we need, and we wouldn't have a good > migration path for RFC production. It's XML, or LaTeX that can be > converted to XML, or something similar. Of course, having a common submission format after the authoring step is a requirement, and XML does that job reasonably well. It is also the format from which I do my extractions. XML has well-understood tools for data modeling (e.g., Relax-NG) that come in handy for defining that submission format. Using a more human-centric authoring language (markdown, LaTex if that rocks your boat) and then translating to XML does introduce the occasional error condition where you actually have to look at the XML to figure it out. Still, the majority of the work can be done in a significantly more humane environment. That is a net win for most people, and it begins to really shine for complex documents and complex collaborations. On the comments about stability: For a decade, RFCXMLv2 has had the advantage that it was relatively stable (defined by a single implementation that for most of that period wasn’t hacked much). That advantage is not available for RFCXMLv3. Authoring formats such as markdown might be instrumental in bridging the period of instability. Grüße, Carsten
