Re: [rfc-i] standards for references/URLs in RFCs ? (Was: Re: archiving outlinks in RFCs)

Jean Mahoney <> Wed, 26 April 2023 19:29 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id 56963C15153F; Wed, 26 Apr 2023 12:29:09 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -4.199
X-Spam-Status: No, score=-4.199 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, NICE_REPLY_A=-0.001, RCVD_IN_DNSWL_MED=-2.3, RCVD_IN_ZEN_BLOCKED_OPENDNS=0.001, SPF_PASS=-0.001, URIBL_DBL_BLOCKED_OPENDNS=0.001, URIBL_ZEN_BLOCKED_OPENDNS=0.001] autolearn=ham autolearn_force=no
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id AdrpRBbVKGzP; Wed, 26 Apr 2023 12:29:04 -0700 (PDT)
Received: from ( []) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by (Postfix) with ESMTPS id E00E3C14CF1F; Wed, 26 Apr 2023 12:29:04 -0700 (PDT)
Received: from localhost (localhost []) by (Postfix) with ESMTP id C9C9E424B43F; Wed, 26 Apr 2023 12:29:04 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id Z0cyHWn67v0e; Wed, 26 Apr 2023 12:29:04 -0700 (PDT)
Received: from [] (unknown []) by (Postfix) with ESMTPSA id 8E40F424B42D; Wed, 26 Apr 2023 12:29:04 -0700 (PDT)
Message-ID: <>
Date: Wed, 26 Apr 2023 14:29:03 -0500
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:102.0) Gecko/20100101 Thunderbird/102.10.1
Content-Language: en-US
To: Toerless Eckert <>, Alexis Rossi <>
References: <> <ZEjKrJK/>
From: Jean Mahoney <>
In-Reply-To: <ZEjKrJK/>
Content-Type: text/plain; charset="UTF-8"; format="flowed"
Content-Transfer-Encoding: 8bit
Archived-At: <>
Subject: Re: [rfc-i] standards for references/URLs in RFCs ? (Was: Re: archiving outlinks in RFCs)
X-Mailman-Version: 2.1.39
Precedence: list
List-Id: "A list for discussion of the RFC series and RFC Editor functions." <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Wed, 26 Apr 2023 19:29:09 -0000

Hi Toerless,

On 4/26/23 1:54 AM, Toerless Eckert wrote:
> Alexis,
> Thanks a lot for the initiative, but before having more opinion about it, i would
> like to reconfirm our current policies for new draft->RFC, and what type of
> references/URLs are permitted/required by our process (IETF and RFC-editor).
> For example, in one of my ongoing WG drafts, WG members made the comment that
> a reference to a github location would not be looked upon favorably by RFC editor.
> Aka: inappropriate reference because it is not stable.
> In that particular case, the github file was part of a presentation given at a
> WG meeting in the past, so our plan for this reference is to use the IETF proceeding URL
> for the presentation instead.
> [ Which should be considered to be an eternally stable URL, one would hope, unless we
>    do get IETF LLC? web admins that like in almost all commercial web pages of our industry
>    seem to be run by admins that like to drive users mad by randomnly deleting important
>    reference/history information or just changing URLs for spite or some new wb page tooling
>    that "did not allow us to keep existing URLs" (typical excuse i hear). /rant ]
> Aka: I am totally unclear what type of URLs are or are not seen as
> appropriate today by IETF process/RFC-Editor, and if someone could point me to any
> reference we have (RFC ? ?), that would be lovely.
[JM] The online portion of the style guide [1] shows which URLs to use 
for RFCs, I-Ds, errata reports, and IANA registries.

Other URLs are assessed for stability by the RPC staff (e.g., Does the 
URL work? Does it redirect? Does it link to a personal site? Are there 
any certificate issues?). If we have concerns about stability, we ask 
the authors during AUTH48 if there is a more stable alternative.

> As another example, when defining terms, i sometimes thought it would be appropriate
> to point to wikipedia. But given how the understanding of terms is changing over time,
> wikipedia definitions might be the worst references to use. Just think of all the
> technical terms we where fond of using (e.g.: blacklist) and that are now shunned/deprecated
> by us (not even sure what the right word for the process is ;-) - as one example category
> for this problem. If at all, it seems to me that references to wikipedia could
> really only go to an archived version of a wikipedia page that was used by the authors
> when writing the draft/RFC.
[JM] Dated URLs are preferred for Wikipedia references.

Best regards,


> Cheers
>      Toerless
> On Tue, Apr 25, 2023 at 11:50:24AM -0700, Alexis Rossi wrote:
>> Hi all,
>> I wanted to let the community know about something I’ve been working on. As you might know, one of my previous jobs was running the Wayback Machine, so when I started working with with this collection of RFCs one of my first thoughts was, “I wonder how many broken links are in these RFCs from the past few decades?”
>> In general, the average lifespan of a URL before the content changes or disappears is on the order of 100 days. Fortunately for us, the links used in RFC references seem to be much more stable than that. For instance, so far I’ve only found one broken link in an RFC from the past 6 months [1].
>> Even though we favor these more stable URLs, some of them will eventually change or go 404 and having archival documents with link rot is something we can take steps to avoid in the future.
>> The first thing I wanted to do was just make sure we were archiving these outlinks somewhere. This won’t fix a broken link in the RFC, but at least the resource can be saved elsewhere for someone curious enough to go look (and potentially we could fix links in some version of the RFC in future).
>> The main services that are well qualified for this purpose are <> (run by the Internet Archive) and <> (run by Harvard Law School Library). I chose Archive-It, and when I approached them they offered us an account [2] with enough free data storage for our needs. Yay for non-profits supporting each other!
>> So far I have used Archive-It to:
>> Archive <>, <>, <>, and <> (minus datatracker and the mail archive)
>> There are lots of references to these sites in RFCs, but I also wanted to preserve the contents for their own sake. I plan to revisit these sites once per year.
>> I am avoiding datatracker (except for outlinks from RFCs) because of concern about the extra traffic causing problems for the team that maintains the site.
>> I have not concentrated on archiving the mail archive yet, though I know some of it has been saved incidentally.
>> Archive outlinks from RFCs
>> About once per quarter I’ll grab the outlinks from newly published RFCs and get them crawled.
>> I am also going backwards through the entire series - I’ve started with the most recent RFCs (links are more likely to still be live) and am working my way back in time.
>> There may be more room for improvements here, for example including archived links in RFCs from the start w here appropriate, or potentially defining a way for links to be self-healing in published RFCs.
>> Please let me know if you have ideas or feedback on this.
>> Thanks,
>> Alexis
>> [1] RFC9311 published in September 2022, in Section 11 (Informative References) this link is 404: <>[2] <>
>> _______________________________________________
>> rfc-interest mailing list