[rfc-i] Comments in the canonical RFCs

Paul Hoffman paul.hoffman at vpnc.org
Thu May 8 11:55:30 PDT 2014


On May 8, 2014, at 11:38 AM, Paul Kyzivat <pkyzivat at alum.mit.edu> wrote:

> On 5/8/14 1:44 PM, Paul Hoffman wrote:
>> On May 8, 2014, at 9:54 AM, Paul Kyzivat <pkyzivat at alum.mit.edu> wrote:
>> 
>>> Maybe this was discussed before, but why are XML comments forbidden in the canonical format?
>> 
>> They could easily make the semantics of the RFC really horrible.
>> 
>> <t>The client MUST abort if it sees this signal.
>> <!-- Paul Kyzivat showed where this might not always be a good idea, but the WG
>> had already lost steam and let it go. -->
>> </t>
> 
> I don't see any problem in having a comment like that remain.
> It doesn't change the semantics of anything. But it does provide some insight that might be useful if/when a bis effort begins. And it might be even more useful if somebody is "borrowing" some xml from this rfc while working on something similar.

OK, have to agree to disagree here. If someone said to me "you implemented from the RFC, but not from all the hidden comments", I would take that as a bad reflection on the RFC series.

>>> I often use comments to provide info of relevance to those editing the document. While these may become less useful when the document becomes an RFC, they could be useful again in the future if a bis version of the RFC is to be generated.
>> 
>> Specific examples would be useful here.
> 
> I'd have to spend more time searching than I really want to invest right now.
> 
> In any case, right now there isn't xml available for RFCs. My xml for drafts is often laced with comments to myself or other authors. Most of them can probably be stripped as the thing gets close to completion. I can just imaging that some might be worthwhile to keep.
> 
> I would expect that when editing is being done to turn the draft into an rfc the editor will have discretion whether to retain comments or remove them. I was only objecting to making that removal mandatory.

If these are editing notes, then some might be retained for posterity but removed from the canonical RFC.

For Internet Drafts, editing notes should be <cref>s instead of comments so everyone reviewing the draft can see them.

The RFC Editor will probably add comments or <cref>s to the document while they are editing, particularly for AUTH48. If the RFC Editor was using a version control system, creating a log of those would be easy. To date, that hasn't happened.

--Paul Hoffman


More information about the rfc-interest mailing list