[rfc-i] feedback on draft-hoffman-rfcformat-canon-others-00,
paul.hoffman at vpnc.org
Mon May 21 15:25:03 PDT 2012
On May 21, 2012, at 4:06 PM, Joe Hildebrand wrote:
> On 5/21/12 1:53 PM, "Martin Rex" <mrex at sap.com> wrote:
>> Paul Hoffman wrote:
>>> Brian E Carpenter wrote:
>>>> I would prefer that CR/NL/LF/TAB are allowed but treated as a single space.
>>>> Isn't that how they are treated in *ML anyway?
>>> So, let me drill down a bit here. Why should that be in the *canonical*
>>> version of the RFC, instead of in one of the additional versions provided
>>> by the RFC Editor? That is, what advantage do you see in having a
>>> canonical version with lines wrapped?
>> I would ask a different question: Why should the canonical format
>> be delibertately obfuscated so that it can only be displayed,
>> edited and produced by special tools?
> I believe the canonical version should be the one that the most people are
> going to prefer use in order to decrease the likelihood that format
> translation errors lead to interop problems.
Given that that is impossible to determine, might it be reasonable to instead allow authors to test the translation before and during AUTH48? That is, if the submission format is either HTML or text, the author could check both during AUTH48. Picky automated tools would start popping up quickly, I imagine.
More information about the rfc-interest