Re: [Gendispatch] I-D Action: trust assets, draft-eggert-ietf-and-trust-00.txt

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

Just for the sake of scientific integrity...

I wrote:

> A rough
> count of the number of distinct author names in the pre-5378 RFCs is
> 3810, although that includes some probable duplicates like 'V.Cerf' and
> 'V.G.Cerf'. (I wrote a Python script to compute that number. Run
> against the complete RFC index, it finds 6328 distinct author names.)

That was done quickly. After fixing some bugs and oversights, here are
better numbers, although they don't change the conclusion: too many people
to contact them individually.

Counting RFCs 1 to 5377
5249 RFCs found; 10964 total authors detected; 3590 unique identities.

Counting RFCs 1 to infinity.
9126 RFCs found; 22057 total authors detected; 5944 unique identities.

Note: Minor errors exist in data.
Authors with alternative initials are counted twice.

Regards
    Brian Carpenter

On 23-Oct-22 10:25, Brian E Carpenter wrote:
> On 23-Oct-22 05:02, John R Levine wrote:
>>> is that asset register publicly accessible?
>>
>> It's on the Trust's web site, with a link on the home page.  But if you
>> had trouble finding it, why didn't you ask?
>>
>>>> If you're wondering where the copyright or license info is for all of
>>>> the pre-3978 RFCs, other than a handful of licenses from ISI and a few
>>>> of the authors, there isn't any. As Joel noted, attempting to
>>>> determine the status of early RFCs would be a hugely expensive and
>>>> unproductive exercise in frustration. In a few cases I doubt that
>>>> anything short of a suit in a US federal court would resolve it (look
>>>> at RFC 20) and I hope none of us want to go there.
>>>
>>> I thought that when the Trust was established, it was supposed to
>>> actively approach authors of pre-5378 documents, especially core
>>> Internet specifications, to get the copyrights signed over. Has that
>>> effort been abandoned?
> 
> As far as I recall, the idea of "actively" approaching those several
> thousand authors was abandoned about 10 seconds after it was suggested,
> as completely impractical. It would have involved tracking down a large
> number of people, or in some cases their executors or heirs. A rough
> count of the number of distinct author names in the pre-5378 RFCs is
> 3810, although that includes some probable duplicates like 'V.Cerf' and
> 'V.G.Cerf'. (I wrote a Python script to compute that number. Run
> against the complete RFC index, it finds 6328 distinct author names.)
> 
> We (the original Trustees) did send messages to the obvious places
> asking people to sign the "RFC DOCUMENTS NON-EXCLUSIVE LICENSE"
> and some people did happily sign it, but the response was underwhelming.
> 
> It is absolutely true that not all the signed licenses are listed in the
> asset register; several more than the three listed were signed in 2007, but
> I have no records. I signed one myself. As a Trustee, I assumed that the
> IAD would retain and archive them. They may be gathering dust in a
> filing cabinet somewhere.
> 
> Overall, the pre-5378 boilerplate proved to be the only practicable
> solution, which puts the responsibility back on the (hypothetical)
> person wishing to make a derivative work. That scales, and I think
> it remains the only practical solution.
> 
>       Brian
> 
>> I wasn't around in 2006, but if you had asked I would have told you that
>> last year while I was a trustee I got licenses from two or three of the
>> early authors who I knew.  Those are the ones that aren't in the asset
>> register yet.
>>
>> But really, it's not that simple.  Anything published in the US before
>> 1976 without a copyright notice is in the public domain.  In many cases it
>> is unclear whether an author wrote an RFC as an individual, in which case
>> he can license the rights, or as an employee in which case whoever he was
>> working for at the time owns the rights, unless the employer was the US
>> government in which case it's public domain.  In many cases the employers
>> no longer exist, and I am unware of a practical way to figure out who
>> might have inherited their assets.  In a few cases (RFC20) it is not clear
>> whether the author even had the rights in the first place.
>>
>> And of course, some authors are dead.  We believe that Jon Postel wrote
>> all his RFCs as an ISI employee, and we have a license from ISI, so we're
>> OK there.
>>
>> Once again, if this was important, why didn't you ask?
>>
>>> I'm sorry if the tone comes across as adversarial, that is not the
>>> intent. The intent is to capture the IETF community's expectations of
>>> the Trust.
>>
>> If that is truly the intent, I strongly encourage you to withdraw this
>> error riddled draft and talk to the trustees about all the assertions you
>> made so you can at least get your facts straight.
>>
>> I entirely agree that the trust's early recordkeeping was poor, and I was
>> as surprised as anyone that there's supposed to be a decade old HSM
>> somewhere.  But it is not 2006 any more, and I believe that the current
>> trust is a lot different from what this document implies.
>>
>> Regards,
>> John Levine, [e-mail], Taughannock Networks, Trumansburg NY
>> Please consider the environment before reading this e-mail. https://jl.ly
>>