On Sep 17, 2011, at 5:37 PM, Yaron Sheffer wrote:
Like Keith, I believe we can benefit a lot from users being able to freely
annotate RFCs with implementation notes, corrections and even opinions ("this
protocol option sucks!").
But I also tend to agree with Joel that the wiki format is inappropriate for
this purpose, because if people are allowed to change one other's comments,
we are very likely to repeat the WG discussion, but without the processes and
incentives that enable us to eventually achieve consensus. So I am not as
optimistic as Keith about the wiki format leading to a status quo. The same
people who would argue their point forever on a mailing list would just keep
editing and re-editing the wiki page.
Maybe a workable compromise would be to let people re-edit their previous
contributions. Though I'm also about the tendency for large numbers of people
to submit irrelevant material. I think that some sort of moderation might be
in order, which begs the question - who should do the moderation?
Slashdot-style moderation, at least, doesn't seem to work well - it favors
those who comment early rather than those who submit the best comments. What
you want to do is favor the contributions that summarize an issue and/or its
resolution fairly, clearly, and succinctly; and then make the set of comments
that do this, and cover the more important issues associated with an RFC, the
ones that are the most visible.
I think the Annotated CPAN example ( http://www.annocpan.org/) is near
perfect for our needs:
- The main text is visually distinguished from the annotations.
- Annotations are visually near the relevant text, rather than appended at
the end.
- The main text cannot be changed.
- Annotations can be freely added by anybody, it is trivial to open an
account.
- Users are identified but with no strong authentication.
- One user can comment on another user's comment, but cannot change it.
- There is some moderation behind the scenes (I haven't studied it, but it's
essential in order to avoid spam).
- There's an average of 1-2 comments a day, so a small number of moderators
can handle the traffic.
As a starting point, this might not be too bad, especially if the code is
available and can be adapted (even if it is written in Perl, sigh). Though I
think that anything of this nature is going to have to adapt over time as
experience with it is gained.
Keith
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf