ID-nits [was Re: suggestions (voting)]

[Dean Willis]

On Tue, 2002-11-26 at 14:46, Marshall Rose wrote:
> > My fantasy is that I could run a draft through a single tool that does
> > all the automateable tests and sends me a report. I don't want to have
> > to, by hand, extract the MIB from a document, clean up all the damnable
> > ASCII page formating, and process it through smilint, then extract the
> > bibliography, clean it up, and process that through reflint, and so on.
> > 
> > How about a nice, simple, tagged input format that has semantic meanings
> > like "This is a mib, check it with smilint" and "this is a normative
> > reference, validate that the URL resolves".
> 
> so, basically, you'd like something added to xml2rfc so that when it
> encounters <artwork/> with a particular type that it runs a program
> to do the validation?
>     

That might work. Or maybe one could extend the DTD (schema maybe) to
actually define a MIB section, and the renderers to pretty-print it.

I think we're actually leading up to having something like a formal
specification language that is partially verifiable as THE submission
mechanism for IDs.

--
Dean