August Derleth said:
Why would you want to cover POSIX when comp.unix.programmer does such a
good job of that already?
However, there may be useful discussion involving libraries which
provide POSIX compatibilities to otherwise non-compliant OSes. I'm not
convinced there's a reason to explicitly recommend against discussing
these sorts of libraries because of their functionality.
Besides, not all POSIX stuff will be on-topic. POSIX defines the standard
command line syntax and some of the standard file locations, among other
things not related to C at all. comp.unix.programmer knows about this
stuff; will you? Will anyone in your group?
Thus the recommendation that system-specific queries go to system
specific groups is the good way to go. If posters follow that, great.
If they don't, they reap the reward of someone telling them they'd be
better off asking in whichever is more appropriate.
General chat devolves into pure noise, and knowledgeable people tend to
dislike pure noise. The comp.* hierarchy is usually held to a higher
standard than alt.* or soc.*.
Wasn't quite meaning alt.* or soc.* type "general chat" there.
I think comp.lang.c is is the way it is for a reason, and you should
understand it before you try to make your own newsgroup.
What works with one group does not work for all. Ideally, the plan is
for the clc purists to keep clc. People who use libraries and want to
discuss getting things to work in wierd and wacky environments and
across disparate platforms, use clcl and perhaps move off to system
specific groups if discussion goes that way.
Some on clc would never be seen on clcl. I'm also of the opinion that
there may be some on clc who will welcome the chance to jump ship to a
group where every fourth article isn't someone saying "not strictly on
topic, go away" in response to questions relating to C programming
.
Again, most people who know about a specific system's stuff will most
likely be in newsgroups devoted to that system, not in comp.lang.c.libs.
Getting knowledgeable people to devote time to a new group will be hard.
By all means, people can be refered to other groups if their discussion
will be better served there.
All APIs are just libraries, at least as far as the userland program is
concerned. This is one reason I think clc.libs will fail: It will attract
a host of system-specific questions but not necessarily any
system-specific expertise.
System specific expertise isn't required. It's not a strictly system
specific group. Want to talk about GTK? clcl. Programming with ncurses?
clcl. Which is a good cross-platform GUI library? clcl.
Third-party canceling is, in my view, Bad and Wrong. It opens the
floodgates for abuse on a simply unprecedented level. I don't want to
subscribe to servers that honor such beasts, and I certainly don't want to
read groups where it's explicitly welcomed.
The BI is strictly about repetitive postings (spam) and overly large
crossposts (spam and pests). That's pretty much it.
The BI is well and good for spam-control, but I think it should be up to
server admins to do it, not vigilantes with explicit liberty to cancel
posts at random.
Not at random.
Q: How does the Breidbart Index (BI) work?
A: The BI is defined as the sum of the square roots of n (where n is
the number of newsgroups each copy was posted to, over a 45 day
period). The standard BI threshold in protected groups on Usenet is
20. The people in the Netherlands nl.* hierarchy have apparently gone
with 8.
Example: If two copies of a posting are made, one to 9 groups,
and one to 16, the BI index is sqrt(9)+sqrt(16) = 3+4 = 7.
Example: If three copies of a posting are made, two to one group
each and the third to 3 groups, the BI index is sqrt(1)+sqrt(1)
+sqrt(3) = 1+1+1.73 = 3.73.
Example: One posting is cross-posted to 9 groups, the BI index is
sqrt(9) = 3 = 3.
Many news servers will have already culled articles cross-posted to 8
or more newsgroups just on principle
. Recommending a BI threshold
of 3 is a very aggressive anti-spam posture. The calculation of BI
levels is discussed in more detail at
http://www.stopspam.org/usenet/
mmf/breidbart.html
Genuine postings are never going to be hit by the Breidbart Index at 3.
Now clcl could easily adopt a less aggressive anti-spam posture to
aus.business if people really want. It'd be embarrassing, but hey. I'm
ready to be convinced by the hoards
.
The server admins feel the financial pain from spam,
since they own the machines spam pollutes most heavily, and they should
be given the reigns when it comes to controlling it.
There's nothing stopping them from keeping the reigns. With most news
server software, you can happily ignore cancels and even block reposts.
Can't say any of that's been relevant or needed for aus.business up to
this point or for the Netherlands nl.* hierarchy, to my knowledge.
I'd think it would be a good idea if I wasn't convinced of its failure.
Not exactly a vote of confidence, but you'll hear worse if this isn't
completely ignored.
Thanks for the feedback.