Re: An IETF repository for working code in our protocols?

[Vijay Gurbani <[e-mail]>]

Dear Joel: Thank you for your time and your thoughts.  Please see inline.

On Wed, Aug 19, 2020 at 11:19 AM Joel M. Halpern <[e-mail]>
wrote:

> I would be very concerned about the IETF appearing to endorse particular
> implementations by archiving the source.  I would be much more comfortable
> with Adrian's suggestion of pointers to implementations that wish to be
> listed (open source or otherwise).
>

I believe that there is an implicit assumption that by archiving the source
code, the IETF is endorsing a particular implementation.  I can see reasons
why it may be construed as such, and perhaps this was what Adrian meant by "the
IETF becoming a venue for advertisements".  This is definitely not the
intended consequence of this process.  I can also foresee that if there is
a de-facto state-of-art archived implementation of a certain protocol, then
derivative implementations may become genetic variations of the de-facto
implementation, with all of the problems that this will entail.  So, yes,
this is a concern, and the question is whether we can mitigate it in some
manner?


> In terms of your concern about wiki maintenance, archives of source are in
> my opinion worse, since they will not have any fixes made after the archive.
>

Right; we cannot mandate the fixes, of course.  But the purpose of such a
repository would be to provide bootstrap code, and allow implementers to
find efficient ways around what they started off as.

In terms of visibility, I do not see why an archive is better than a  wiki
> page.  Neither should be called out in the RFC, since they are transient.
>

I understand and respect that point of view.  But I will note that RFCs
themselves are transient.  Even an Internet Standard can be made obsolete
at some time by a new RFC-to-be.  So I don't see the above as problematic,
and instead see that it benefits implementers who are not part of the IETF
lore and just look at the RFC to figure things out.  (And I suspect these
implementers are far more numerous than those of us who attend the IETF and
write the protocols.  My interest is in reaching out to these implementers.)

Thanks,

- vijay