Things that used to be clear (was Re: Evolving Documents (nee "Living Documents") side meeting at IETF105.)

[Andrew Sullivan <[e-mail]>]

Hi,

I work for ISOC but don't represent it in this discussion.  Also, I
seem to be sending more mail than is my habit today.

On Wed, Jul 03, 2019 at 07:21:26PM -0400, Keith Moore wrote:

> difficulties.    It used to be clear that you didn't deploy implementations
> based on Proposed Standard, but people did anyway.

When was that "clear"?  Just to pick a random example, RFC 2181
includes clarifications to the DNS specification that are absolutely
critical to the way DNS actually works, and that are certainly widely
deployed.  It's arguably something that ought to be considered a
formal part of STD 13.  It was published in 1997 and is a Proposed
Standard.  Given that STD 13 is only 10 years older than that, we've
lived with this Proposed Standard for much longer than we lived with
the unclear bits 2181 was designed to address.

And, of course, when STD 13 was published as RFCs 1034 and 1035 back
in 1987, it _wasn't_ an Internet standard, yet it seemed to get
deployed.  Indeed, the whole idea of "Internet Standard" seems to be
rather fluid leading up to the publication of RFC 1310, which came out
in 1992.  I note that RFC 1011, from 1987, lists the DNS with
"Recommended" status, but does not refer to the protocol documents we
rely on today but instead refers to RFCs 881-883.

I'm not trying to be querulous.  But some of my recent experiences
make me leery of assertions about how things got deployed in the past,
because I don't think that things were nearly as tidy as many people
today seem to be suggesting they were.  I was not among the people
connected to the Internet in the 1980s, but I know when I connected in
the 1990s I was both surprised how well things worked and dismayed at
how poorly they worked, often at the same time.  That experience, at
least, has not changed!

Best regards,

A


-- 
Andrew Sullivan
[e-mail]