[rfc-i] Five-author maximum?
braden at ISI.EDU
Thu Jun 10 11:42:16 PDT 2004
I would like to make some general observations about the author list
question, from our viewpoint here in the RFC Editor organization.
We initiated the limit on authors, maybe a year ago (? ... how
time flies), when we observed a disturbing trend towards RFC author
One factor driving this inflation seemed to be in increasing
involvement of players from industry, particularly the telecomm
industry, on the IETF stage. There seemed to be an urge for every
company to get their name on RFCs, by having one of their guys as one
of the authors, and the author lists were beginning to read like an
article on the telecomm or IT industry in the WSJ. This seemed to us
to be contrary to the long-standing tradition of individual, not
corporate, efforts in the Internet technical community.
The other tendency was to include everyone in a WG who contributed
to a document. But this is a slippery slope, and we seemed to be
starting a serious slide down it.
It is true that this is not a new issue. In 1987-89 I chaired the Host
Requirements Working Group. The Acknowledgments section of RFC 1122
lists some 25 major contributors, and another 25 who contributed some.
In fact, there were probably 4 or 5 people who contributed significant
nibbles of text that got incorporated, but it would have been really
tough to draw the line. Fortunately, the WG felt comfortable with
listing only the Editor on the document, so I did not have to make
the judgment on who should be listed (clearly, not 25!)
Authorship can of course be a minefield of personal anxiety. Some
people care a great deal whether they are included, sometimes even
though their contribution has been marginal. A futher problem arises
in those cases (perhaps few, these days) when the starting point for an
IETF spec was an earlier research effort. Should all those who
contributed to the earlier research, but did not help write the
subsequent IETF spec, claim authorship on the final RFC? It seems
dubious to me, but I have received anguished messages on the topic. It
seems that this could logically lead to a BGP4 RFC that carried the
names of everyone who had contributed to BGP development over the
We also note that some academic fields have a tradition of including
everyone remotely involved. You not infrequently see articles in
Physical Review Letters with 20 to 50 authors. OK, that's their thing,
but that has not been the way of the Internet/IETF.
We therefore proposed an author list limitation to the IESG and to the
community. To help take the sting out, we invented a new official RFC
section, the Contributors section. Fortunately, in fact, nearly
everyone has accepted the general principle and cooperated in good
spirit, and we are the RFC Editor are grateful. The (admittedly
slightly arbitrary) limit of 5 has generally worked out. The limit
gives WG chairs a stick that they can (and do ;-)) wield to control
their author lists.
But life is full of exceptions, and Jon Postel taught us to maintain
general principles while being willing to Do The Right Thing in special
cases. We try to maintain that tradition of principled flexiblity.
There have been a (very) few exceptions to the rule-of-5 for author
lists, but only after really good arguments were made. The topic of
this discussion is the most recent example.
Bob Braden for the RFC Editor
More information about the rfc-interest