tower@wheat-chex.ai.mit.edu (Leonard H. Tower Jr.) (03/15/91)
[ this article has been posted directly to comp.lang.c++.
it was not gated from help-g++. -len
]
It's been decided to leave the gateway up. For these reasons:
- the vote was poorly conducted by USENET standards. For
example, there was no announced discussion period before the voting
started. And the discussion before the vote was mostly opposed, while
that after the voting started was mostly in favor.
- Michael Tiemann, the creator and maintainer of g++ believes
that the gateway is useful and should stay up. He didn't hear of the
vote until after most of the votes had been cast. I've appended the
posting he made at that time.
- the number of people who voted compared to the number who
read the group is very small.
- the margin of yays over nays was small, and well under the
100 difference usually used for votes deciding USENET group issues.
The actual vote report from Karl Kleinpaste is attached.
- Several people have told me they still get only
comp.lang.c++ at their site, and not gnu.g++.help. Though they could
switch over to getting it via mail, they don't wish to.
- Since the vote was called for, articles gated from help-g++
has been given a distinctive Organization: header line:
"Organization: Gatewayed from the GNU Project mailing list help-g++@prep.ai.mit.edu"
This allows people who don't wish to read these articles to
add this line to their rn kill files (or the equivalent in other news
readers). This allows everyone to have what they want, so it doesn't
seem appropriate to force the will of a majority onto a minority.
Even so, we are uneasy about not honoring the results of the vote.
But as each reader can now vote no in his kill file, it seems best to
leave the gateway up, so everyone can have what they want.
I appreciate comments to me as:
tower@prep.ai.mit.edu
instead of posting them to comp.lang.c++ (well ok, expect one from
Jim Adcock :-< ). If there is enough dissatisfaction with not
honoring the vote, we'll conduct another one using the USENET
standards. Though I really hope the matter can be laid to rest now.
thanx -len
----------------------------------------------------------------------
Return-Path: <tiemann@eng.sun.com>
Date: Wed, 5 Dec 90 09:30:04 PST
From: tiemann@eng.sun.com (Michael Tiemann)
To: lang-c++-netnews-dist@prep.ai.mit.edu
Cc: help-g++@prep.ai.mit.edu, rms@wheaties.ai.mit.edu
In-Reply-To: Jim ADCOCK's message of 4 Dec 90 18:42:34 GMT <59548@microsoft.UUCP>
Subject: Anyone for decoupling GNU and c.l.c++?
Reply-To: tiemann@cygnus.com
Organization: Cygnus Support (+1 415 322 3811)
There is a reason why the gnu.g++ and comp.lang.c++ newsgroups were
linked. It was not, as Jim ADCOCK suggests, because one or the other
was not strong enough to stand on its own readership. Rather, it was
because many people (a vast majority of those who voted) wanted to
share experiences and information between the two groups. Here are
some examples where multiple inheritance (for newsgroups) is useful:
1. A naive G++ user has a question about overloading operator+.
Should s/he ask in the g++ forum or the comp.lang.c++ forum?
Usually the answer is unclear, so people cross-post. As a
matter of fact, one of the reasons that these groups were merged
was because there was so much cross-posting going on.
2. An unspecified C++ user finds that cfront and g++ have different
behavior for a complicated inheritance lattice. Both results
are unexpected (this has happened). Where does s/he post?
3. A C++ expert has just released a cool, freely redistributable
C++ interface to X/NeWS. Where should the announcement go?
4. A bored graduate student hears about the above library, but
doesn't know where to get it. To which group should he pose his
request for information?
The list goes on and on. Everybody has their own quirks about what
they will and will not tolerate about people's postings. Some people
get upset when they see postings about how clever they were at
painting themselves into a 640K address space. Some people are
dismayed by other's blithe acceptance of UNIX.
If it absolutely ruins your day that somebody doesn't understand how
to overload operator+, how segmented architectures are really better
than large linear address spaces, or how Apple and Lotus are taking
away programmer's freedoms, I believe that there is a newsreading
service which, for a small fee, will delete these messages before you
have a chance to read them, and you need never know they were ever
written.
In the mean time, please don't ask to have everybody's service
disconnected because you don't like yours.
Michael
----------------------------------------------------------------------
From: karl_kleinpaste@cis.ohio-state.edu
Date: Tue, 5 Feb 91 10:44:17 -0500
Subject: Survey results
I just realized I'd never summarized the responses I'd gotten to the
survey over whether to perpetuate the double one-way gate of help-g++
<=> comp.lang.c++,gnu.g++.help.
The short version is:
14 no including
no, but avoid politics.
no, just got off help-g++ because i can read comp.lang.c++.
no, it's not exactly swamping us.
36 yes including
yes, if politics continue.
yes; admirable experiment which failed due to, 1, incompatible charter,
2, one-sided discussion.
"No" meant "keep the double gate" while "yes" meant "drop it."
It's about 5:2 for dropping it. But it may not matter any more.
--karl
----------------------------------------------------------------------
Len Tower, Project GNU of the Free Software Foundation
1000 Mass. Ave., Cambridge, MA 02138, USA +1 (617) 876-3296
HOME: 36 Porter Street, Somerville, MA 02143, USA +1 (617) 623-7739
UUCP: {}!mit-eddie!mit-prep!tower INTERNET: tower@prep.ai.mit.edu