RFC - e-mail in tough environments

Yavor Doganov yavor at gnu.org
Thu Dec 10 17:35:26 UTC 2009

Matthias Kirschner wrote:
>     - *E-Mail signature* Keep it small and simple. Signatures longer
>       than five lines should be avoided.

IIRC in the good old days the standard was maximum 4 lines (5 lines
with the "-- " sig separator).

>     - *Mailinglists* Use list-reply. It is not necessary to include the
>       sender in To: or Cc: if he is subscribed. If the e-mail programs
>       are configured correct the sender will be Cc’ed if he is not
>       subscribed or wishes to be Cc’ed.

Hmm.  This is certainly debatable.  On GNU and GNOME mailing lists
(and I dare to say, on *most* free software lists) the default policy
is to always to CC people who participated in the thread.

This is fairly logical:

Scenario 1)

  I am a user who discovered a bug in GNU Foo, and post to
  bug-foo at gnu.org.  I should not be required to subscribe, and I
  should not be required to dig eventual followups via inconvenient
  interfaces like Mailman's archives, or -- heaven forbid -- something
  like the various/nefarious mail archiving sites like nabble.com.  I
  should get the response right away.

Scenario 2)

  I'm a regular lurker on the bug-foo list, thus I'm subscribed.  A
  user reports a bug/misfeature, or is confused about something in the
  `foo' package, so I reply to him.  It is not feasible to check
  whether the user is subscribed, even if I happen to be the list
  admin.  I just CC her, to be on the safe side and save her extra

Scenario 3)

  I happen to develop gnome-foo and need to announce a string change
  to gnome-i18n during a string freeze.  I'm not a translator myself,
  and I'm not subscribed to that list.  I post the announcement, but a
  translator then asks a question about the new (cryptic) string,
  which I don't receive.

Scenario 4)

  I'm subscribed to many lists, but I don't read regularly all of
  them.  However, if I participate in a discussion on some of those
  "low interest" lists, I appreciate if a person who posts to the
  thread CCs me.  I know this is not very usual situation, but the
  more mail I'm processing as years go by, the more I appreciate it.

Scenario N)

  There are plenty.  Really.

From all the lists I'm subscribed to (much more than 200, actually),
TTBOMK only Debian has a strict no-CC policy.  I comply, of course,
although I notice that many DDs don't bother.  (The extra mental
excercise to determine where you're posting to is also slightly
annoying.)  In general, it seems to me that a no-CC policy is very
inter-community friendly (you sort your mail easily, and you rely
people to CC you when you ask), but basically user-unfriendly -- it is
very presumptuous to rely that the OP is going to search and watch for
your-almighty-followup just because you happen to have an odd list

More generally speaking, mailing lists are a gross hack to replace
Usenet newsgroups, just because the new kids on the block seem to find
NNTP archaic and weird (partly because most "modern" user-agents have
from moderate to poor NNTP support).  Thus the always-CC vs. no-CC
debate.  YMMV.

More information about the Discussion mailing list