John,
Is there any reason to not publish the v18 draft as an informational RFC, since it's in use in the field, and then supersede it when the final RFC is published?
Cheers,
Andy
On Wed, Jan 12, 2022 at 7:23 PM John Kunze <jak@xxxxxxxx> wrote:
Thanks for CCing me, John. I can see how this ARK spec evolution looks a bit unusual, especially given its 20-year run. It's taken a while to get real experience with persistent identifiers.In fact this run of drafts is heading towards a projected submission as an Informational RFC of April 2022. Things got somewhat reactivated in 2019 when the "ARK Alliance" (arks.org) established a technical working group that's been shepherding along a set of changes and updates based on deployment (of an estimated 4.2 billions ARKs by some 880 institutions).The very latest flurry of revisions is me thrashing around trying to figure out how to use the martinthomson/i-d-template github repo and setting things up as a github organization. Please pardon my noise (I started this spec using nroff ;).I can't comment on how others may have been (ab)using the internet draft system since I only know it through my keyhole view. Personally, I love it and frequently cite it as a very fine example of the open internet and its values. There were two "long-term stable" drafts in 2008 and 2018, which I did manage to separately archive because I didn't know how long the expired drafts would be available.In the "Transitional Note", would it be better for us to refer to the 2018 (v18) draft via the archived copy's URL rather than via the datatracker copy? Perhaps that would be more appropriate.I'm quite pleased that everyone's old drafts do remain available since as a researcher/scholar (not as an IPR lawyer) I'm often trying to find out the details of an idea that I vaguely recall even if the ideas didn't get developed further (at least not in I-D form). I think there's lots of valuable history in the IETF repository and the tools make the whole corpus easy to use.On Wed, Jan 12, 2022 at 3:10 PM John C Klensin <john-ietf@xxxxxxx> wrote:Mike,
First, to what Scott and Fred said, yes.
Prior to when we made them readily available after the nominal
six months, they were still available to anyone who showed up
with a subpoena or, IIR, a small wad of cash to cover presumed
Secretariat expenses to dig them out, but my recollection (which
Scott can probably confirm and recalibrate) is that just became
too much trouble... for us and the Secretariat, not (just) the
lawyers.
Partially for the reasons you identify (which may not apply to
this draft -- see below) I've wondered whether we could make
drafts moderately easy for the lawyers to identify and find but
destabilize URLs pointing to them. For example, instead of
having expired drafts with stable URLs, we could move them to
this-year.datatracker.ietf.org, then move then again to
last-year.datatracker.ietf.org, and so on, maybe making them
still possible to find via the Data Tracker, but making stable
references to the URLs impossible and getting to them a bit
difficult or annoying, but not terribly so. Or, as a different
example, we could get rid of the equivalent to a stable
https://datatracker.ietf.org/doc/draft-whatever-whateverElse/
and, after six months, change the file names to
draft-screech-whatever and set the tracker up so that, instead
of supplying a clickable-URL, we simply supplied the equivalent
of "screech-whatever" as text and made people copy that string
and paste it in somewhere. The obvious question is whether the
tooling to support such a scheme would be worth the trouble.
My impression is that, in the past, the answer has been "no"
but, as we review and rewrite various tools, that decision may
be worth reviewing.
That said, this document, however evolved, may be a poor example
of your concern. We do so say (see the first paragraph of
https://www.ietf.org/standards/ids/ ) "Other groups may also
distribute working documents as Internet-Drafts." so we are
explicitly authorizing the behavior with which you are
concerned. And John Kunze (copied) is a long-time and very
constructive IETF participant -- see RFCs including 1625, 1736,
2056, 2413, 2731, 5013, 5791, 8493, and 8574 -- who perfectly
well knows what the rules are. If I had to make a somewhat
educated guess (I hope he will confirm or correct as needed),
what is going on is exactly what the Abstract says is going on,
i.e., that this is a gradually developing and evolving spec in a
different group (remember, we said such groups could use I-Ds
for that purpose), that it is valuable to both them and us to
have the IETF community be aware of the work, and, I hope that,
when it is stable, they will either bring it to us for
standardization or bring a stable and approved version to the
IETF or ISE for informational publication as a standard from
another SDO.
The ARK Identifier scheme is another variation on the idea,
mentioned in RFC 3986, of document identifiers/ names rather
than locators. I think or at least hope, that we gave up on the
"one name system to rule them all" theme. More familiar
variations on that theme (with different design assumptions and
success criteria) are URNs (RFC 8141) and DOIs.
John (Kunze), anything to add?
best,
john (klensin)
--On Wednesday, January 12, 2022 13:24 -0800 Fred Baker
<fredbaker.ietf@xxxxxxxxx> wrote:
> The primary reason that drafts remain available after six
> months is, I'm told, that lawyers want to be able to look up
> IPR. I suspect that your suggestion, which I otherwise
> support, would run afoul of that. What might be good would be
> to describe a mechanism that would allow the draft to vanish
> in six months but somehow remain available to an IPR search.
--On Wednesday, January 12, 2022 16:20 -0500 "Scott O. Bradner"
<sob@xxxxxxxxx> wrote:
> leave it be
>
> having the old versions accessible is very useful for IPR
> disputes
>
> and I'm not sure we could write a ruleset that said no 3rd
> party standards repository that would not impact our own work
>
> Scott
>
>> On Jan 12, 2022, at 4:12 PM, Michael StJohns
>> <mstjohns@xxxxxxxxxxx> wrote:
>>
>> Hi -
>> I got curious when I saw a draft being published with a
>> version number of -33 -
>> https://datatracker.ietf.org/doc/html/draft-kunze-ark-33. I
>> got even more curious when I noticed that -00 of that ID had
>> been uploaded back in 2001. So I read the intro in the -33
>> version and found this:
>>
>>
>>
>>> This is a transitional draft. The ARK Alliance Technical
>>> Working Group (
>>> https://wiki.lyrasis
>>> .org/display/ARKs/Technical+Working+Group)
>>> is in the process of revising the ARK spec via a series
>>> of Internet- Drafts. While the spec is in transition,
>>> new implementors should follow
>>> https://datatracker.ietf.org/doc/html/draft-kunze-ark-18.
>>
>> -18 having been uploaded in 2013.
>> I'm not sure there's anything that can or should be done, but
>> IDs are supposed to be transient documents that either go
>> away or lead to an RFC. Looking at the update history for
>> this document, I'm pretty sure the draft system has been
>> co-opted to be a standards repository for this specification.
>> AFAICT, this draft has never been submitted - in 20 years! -
>> to any RFC stream for publication and that's at least a
>> violation of the spirit of the ID system. I.e. a violation
>> of:
>>
>>
>>> It is inappropriate to use Internet-Drafts as reference
>>> material or to cite them other than as "work in progress.
>>>
>>
>> Perhaps we may want to think about making URL references to
>> previous (or long expired) versions quite a bit less stable
>> to avoid gaming of the system like this? Or prohibit updates
>> of draft chains once a draft has been expired for a year?
>>
>> Or leave it be?
>>
>> Mike
>>
>>
>>
>>
>
