[rfc-i] Private documents [was Alternatives to 'deprecated' in xml2rfc v3]

Paul Hoffman paul.hoffman at vpnc.org
Wed May 7 17:09:59 PDT 2014


On May 7, 2014, at 5:04 PM, Nico Williams <nico at cryptonector.com> wrote:

> On Wed, May 7, 2014 at 3:35 PM, Brian E Carpenter
> <brian.e.carpenter at gmail.com> wrote:
>> Yes, this approach is clean and meets the law of least astonishment, I
> 
> I think so too :)
> 
> More than that, it's also general in just the right way without adding
> too much verbosity or undue burden on tooling.
> 
>> think. I'm not sure if 'standard' is the best attribute name, but
>> only because <boilerplate standard='none'/> looks a bit odd.
> 
> - if <boilerplate> is absent -> standard boilerplate for IETF
> 
> - if <boilerplate> is present but empty (i.e., lacks child text
> nodes), and lacks the standard (or whatever) attribute -> no
> boilerplate
> 
> - if <boilerplate> is present and empty, but has a standard (or
> whatever) attribute -> you get the named boilerplate unless it doesn't
> exist / isn't known, in which case you get an error
> 
> - if <boilerplate> is present and non-empty you get that text as the
> boilerplate (or an error if the standard attribute is also set?)
> 
> Possible attributes:
> 
> - standard -> value names standard boilerplate
> - link -> value is the URI of standard boilerplate
> 
> Anything else?
> 
> Any remaining objections?

Yes: this is much more complicated than it needs to be. Basically, this started with a request for private document processing to not be broken. It then careened into "if we ever want X, we need to be able to specify it in the document". At no point did anyone suggest that "I want to write my own boilerplate text" was anything other than a far edge case.

--Paul Hoffman


More information about the rfc-interest mailing list