Re: [Tools-discuss] Listing updating RFCs in RFCs

[Brian E Carpenter <[e-mail]>]

One more comment in line:

On 09-Apr-24 19:48, Tim Chown wrote:
> Hi,
> 
>> On 8 Apr 2024, at 22:11, Brian E Carpenter <[e-mail]> wrote:
>>
>> I think this discussion belongs as much on rfc-interest as here.
> 
> Feel free to send it there :)
> 
>> In line:
>>
>> On 09-Apr-24 02:55, Tim Chown wrote:
>>> Hi Bob,
>>>> On 8 Apr 2024, at 15:48, Bob Hinden <[e-mail]> wrote:
>>>>
>>>> Tim,
>>>>
>>>>> On Apr 8, 2024, at 7:07 AM, Tim Chown <Tim.Chown=[e-mail]> wrote:
>>>>>
>>>>> Hi,
>>>>>
>>>>> I was caught out this week looking at RFC 4861 as part of reviewing the draft on a new P bit for PIOs.
>>>>>
>>>>> I missed the “updated by RFC 8425” on RFC 4861 as I was looking at the HTML rendered version of RFC 4861 at
>>>>> https://www.rfc-editor.org/rfc/rfc4861.html
>>
>> That is the RFC Editor's version of an htmlized old-style RFC, and apparently it has a bug. That needs to be reported [e-mail] <mailto:[e-mail]>if you haven't already done so.
> 
> I haven’t, I perhaps rather lazily assumed raising it here would achieve that, but there is the bigger picture of the different formats and what metadata is associated with them and how or if that metadata appears in the document body or not.
> 
>> (I don't know why the RFC Editor and the datatracker don't use the same version of the htmlizer, but I imagine there is a good historical reason.)
> 
> I didn’t realise there was a difference.  I assumed the RFC Editor would use the datatracker as the authoritative source.
> 
>>>>> rather than the datatracker version at
>>>>> https://datatracker.ietf.org/doc/rfc4861/ <https://datatracker.ietf.org/doc/rfc4861/>
>>>>> where you can see RFC 8425 (and RFC 9131) listed as updating RFCs which are missing on the HTML rendered version.  There was a clue I missed in that the 7th of 9 RFCs in the Updated by list had a comma after it, but no RFC.
>>>>>
>>>>> The text and PDF versions don’t show any of the updating RFCs for RFC 4861.
>>
>> No. The plain text version is immutable by definition, and as far as I know the PDFs for all RFCs before RFC8650 are simply images of the plain text version and also immutable.
> 
> Why does it have to be immutable?  The html and htmlized versions change over time.  You could track the version history if desired.

Oh dear, you just almost kicked over a hornet's nest.

Since before I first participated in the IETF, it's been a strict rule of the RFC series that nothing, not even a single comma, is changed in a published RFC. This is so that they can be treated as stable references, so when someone cites RFC1234, everybody knows precisely what that means. The doctrine is: if there's a serious error, publish a new RFC that obsoletes or updates the old one.

The RFC errata system is a sort of patch on that doctrine.

The htmlized versions of plain text RFCs are another patch on that doctrine.

Where available, the "verified errata" versions of RFCs are yet another patch.

Then came the "revolution", so that all RFCs starting with RFC8650 are published as HTML, with the plain text version being a bit of a poor relation. But the immutability doctrine hasn't changed.

And now we have the RFC Series WG (RSWG), which is not an IETF WG, where changes in RFC policy are discussed, and things like (im)mutability, errata, and version numbering are much discussed.

To see the resulting hornets buzzing around, see https://mailarchive.ietf.org/arch/browse/rswg/

    Brian



> 
>>>> The htmlized version:
>>>>
>>>> https://datatracker.ietf.org/doc/html/rfc4861 <https://datatracker.ietf.org/doc/html/rfc4861><https://datatracker.ietf.org/doc/html/rfc4861 <https://datatracker.ietf.org/doc/html/rfc4861>>
>>>>
>>>> Does show the list of updated RFCs in the sidebar.
>>> Thanks, so should we consider the HMTL version as obsolete or no longer supported?
>>> I don’t mind as such, but personally I won’t be looking at the HTML version again having been bitten here. Unfortunately it’s the version google search returns first, so many people will routinely find it.
>>
>> As I said, I don't know why the two htmlized versions are different.
>>
>>> Maybe the HTML version should have a clearer indication that the link to “Tracker” at the top is the authoritative source from which other versions can be found?  Quite a few people less familiar with IETF process may have no idea what “Tracker” means.
>>
>> The *definitive* source for RFCs isn't the datatracker, it'srfc-editor.org <http://rfc-editor.org/>. So they need to fix their bug. However, the correct place to start ishttps://www.rfc-editor.org/info/rfc4861 <https://www.rfc-editor.org/info/rfc4861>.
> 
> Thanks, as above I didn’t know that the RFC Editor version was the real source.
> 
>> Unfortunately we can't control what the search engines find first.
>>
>> (The best source of the metadata ishttps://www.rfc-editor.org/rfc/rfc-index.xml <https://www.rfc-editor.org/rfc/rfc-index.xml>if you have the stomach for it. Or in human form,https://www.rfc-editor.org/rfc/rfc-index.txt <https://www.rfc-editor.org/rfc/rfc-index.txt>)
> 
> I think what I’d hope for is an authoritative (single) source that gives the content and metadata about a draft, and then renditions of that into formats where the rendered versions are consistent.  At the moment they aren’t.  We should probably be considering what people new to th IETf and standards documents make of it, as we (or rather you and John!) know the specific details and history that may not be obvious to such people.
> 
> Worst case, the Updated by bug needs a fix.
> 
> Tim
> 
>>
>>   Brian
>>
>>> Tim
>>>> Bob
>>>>
>>>>>
>>>>> I’m wondering whether the updating RFCs should just be metadata to the RFCs, or be included fully in all RFCs in all versions.  It’s a little dangerous/confusing when you only see a partial list.  Or maybe I shouldn’t be looking at the HTML but rather the HTMLised version?
>>>>>
>>>>> Tim
>>>>> ___________________________________________________________
>>>>> Tools-discuss mailing list - [e-mail] - https://www.ietf.org/mailman/listinfo/tools-discuss
>>>>
>>>> ___________________________________________________________
>>>> Tools-discuss mailing list - [e-mail] - https://www.ietf.org/mailman/listinfo/tools-discuss
>>> ___________________________________________________________
>>> Tools-discuss mailing list [e-mail] <mailto:[e-mail]>-https://www.ietf.org/mailman/listinfo/tools-discuss <https://www.ietf.org/mailman/listinfo/tools-discuss>
>