D3U@PSUVM.BITNET (03/06/85)
If I understand correctly that Lauren's claim is that some of the cost of run-
ning Usenet can be cut out by using mailing lists instead of newsgroups, then I
conclude that keeping the number of articles to a minimum keeps cost to a mini-
mum and also that it requires some cost to keep a file,or exec or whatever, in
the software so that more bits have to be transmitted, which increases cost of
line use. This does seem logical in itself. And I know that phone bills get to
be expensive, from previous postings. Perhaps some account of this cost with
respect to , e.g. cost of maintaining a newsgroup and the files within that
group, could be posted (although I know there is a lot of volunteer work going
into Usenet).
The problem I see with replacing news groups with mailing lists is that the
list is essentially or actually invisible to anyone not on the list. Perhaps we
need a posting of mailing lists. But that posting would itself be huge and ex-
pensive. Problems for efficiency arise when subscribers relocate. Also, each
entry for a posted mailing list should probably have a description of the inten
tion of that mailing group, which would increase the cost again. (Whereas it it
easy to figure out the purpose of most news groups without the description)
As for there being too many news groups:
One can unsuscribe to a given group unless that subscriber's software doesn't
permit it.
Chuq comments that there are too many vaguely defined groups. Maybe. But it is
also the case that many topics overlap. For example there is currently a call
for discussion on starting net.cogsci. There is already net.ai and net.cogeng
and net.philosophy and net.psychology. Cognitive science spans all that, and a
particular article on cog sci could be appropriate for even other groups while
being less appropriate for the above four. Cognitive science is a widely hybrid
field. I would agree that one interested in cog sci would look at all possibly
related groups, but I would also agree that it would be better to have a group
that could catch all the oddball articles. One interested in cog sci may not
want to watch net.phil for the few articles of interest outside of the mind/
body problem. Some software is very inefficient in handling articles- details
further down. Some confusion undoubtedly will always be around if there is a
large crowd with varied interests. Too few groups crowds a wide variety of art-
icles under one heading and forces many articles to be place in semi-appropriat
places. Too many groups can requiremultiple postings to many groups. Currently
Usenet is in the middle somewhere- some articles have to be posted to several
groups while others sort of fit particular groups but really should be in a yet
uncreated group (and there always be a few such articles).
The presence of a news group encourages traffic. For instance net.music is
subdivided, the latest sub group being net.music.synth. It seems that there are
a lot more articles on synthesis now than before the group was created, and it
grew all at once to a well used group. Some rarely used groups such as .poetry
have justification in that although few post to it, it is also true that few
are poets or are brave enough to post their own poems.
Some ambiguity will be in the minds of users who may not be sure what is
going on, such as a new initiate. Of course we expect the initiate to figure
out things with experience. I am sure that the Usenet participants come from a
wide range of types, experience, etc. Some users are longtime users 20 years
into their careers. Others are students not far into college. Some are self-
styled anarchists while others have problems coping with the slightest bit of
anarchy or uncertainty.
Software, hardware problems:
Different systems handle net news differently. I am used to a full screen ter-
minal. I can see up to 20 article titles at once. Then I push a button to clear
the screen for the next 20, which appear in a few seconds. To go to the next
news group I press a button, so I can look at a few hundred article titles in a
short time. If I want to look at only a group or two, I ask for one, look at
the titles, and then ask for the next group. I have no need or desire to unsub
scribe to any group. But I found that a terminal that prints line by line would
quickly make me want to unsubscribe to the groups I thought I'd rarely if ever
read. The line by line system shows each article title seperately, and a key
has to be pressed to read or reject the title. That is inefficient, to say the
least. I mention this to point out that the working conditions are different
for different users. With my system I would be happy for another 50 new groups
even if I wasn't interested in many of them, but I can see how someone else
would shudder at the thought.
Usenet is a service. It reaches a lot of people. Each news group serves
an interest of at least a few individuals. A service must -be- a service to be
successful. Having net.music with a couple sub groups is great. Having the poss
-ibility of starting net.cogsci in addition to .phil, .cogeng, .ai, .LISP, etc.
is great. With all the related interests net.cogsci would satisfy a lot of de-
sires and demands, and would probably be an improvement (although it may be
only a small one and the practicability of implementing it is the deciding fac
tor, including how much more cumbersome that new group along with other new
groups would make Usenet for users with inefficient systems. Perhaps I should
admit that at first I thought all the talk about too many groups was silliness,
pure and simple, from people who mostly cared only for their own special in-
terests until I found that I am lucky to work with an efficient system. So I
can afford to not be bothered by inappropriate placement of articles, excessive
duplication, etc.) Having such a resource as Usenet in its current form with
many newsgroups is great. It is a nice service as it is, even if improvements
could be made for different difficulties different users may have (I tend to
think that better software and hardware programming is the best answer to most
of those problems).
CONCLUSION: I tend to think that Lauren is acting in a paranoid manner
when he says lotsa new groups will kill Usenet before its time. If a new group
proves to be used by only one or two individuals, it can probably be dropped.
But I do agree that some discretion must be maintained and that related groups
could be brought together in some cases. e.g. net.cogeng, and net.ai could be
brought together with net.cogsci as subgroups (OR give all related groups
pointers so that when one asks for RELATEDSUBJECTS AI, .cog eng and .cogsci and
.LISP would be displayed and .psychology and .philosophy would be listed as
secondary related groups. This would require more software and hacking, and is
another layer of complication, but it may help with the problem of information
classification). I think Usenet as a bulletin board is nowhere near it's point
of incompetency, although I have no idea how costs figure in- I suspect that
they do not easily permit a full line bulletin board without considerable ex-
pense, and it is in this respect that Lauren may be right.
(perhaps I should admit that my first impressions were that Lauren is
one of those who has a vast preference for order and can't use rules as guides
rather than always as sacred absolutes, and so is happy only where there is the
most obvious and guaranteed order. But he is probably forced into some con-
straints arising from practicalities. I do not mean to make a judgement on him,
other than to say that I think he does spend a lot of time working on problems
of the Usenet and Stargate, which some are complex and demanding. Any other
judgement, or almost any other whether good, bad, indifferent, would be from a
very uninformed position. I don't know how far to agree with him on problems of
Usenet or Stargate, or how far I should disagree.)
I do not intend this as the last words, or even my last words on the sub
ject. I am posting this as my interpretation of what I see going on and I look
forward to response and discussion to help clarify things more.
Please forgive all the interjections that may sound overly personal.I feel
a little self conscious about posting such remarks. I am making an attempt for
truth and fairness.
------------------------------------------- Raychuqui@nsc.UUCP (Chuq Von Rospach) (03/09/85)
In article <1631D3U@PSUVM> D3U@PSUVM.BITNET writes: > The problem I see with replacing news groups with mailing lists is that the >list is essentially or actually invisible to anyone not on the list. Perhaps we >need a posting of mailing lists. I should point out that I am maintaining a list of known mailing lists available on the uucp network, and post it to mod.newslists about every two weeks. People can track down mailing lists through that posting and ask the contact to add them to it. >Chuq comments that there are too many vaguely defined groups. Maybe. But it is >also the case that many topics overlap. For example there is currently a call >for discussion on starting net.cogsci. There is already net.ai and net.cogeng >and net.philosophy and net.psychology. Ambiguous topic names are our single largest problem on the net right now in my eyes. There are simply too many places for many topics to show up, and none of them are 'best' for it. Some places are simply misnames-- net.nlang is an example-- I don't really remember why all of the subgroups that should be in net.culture (net.culture.greek, etc) are in net.nlang, but they really have nothing to do with natural language. How about net.usoft? Nmemonic, no? (Its about microsoft software, obviously-- but that's wrong, too!). What we've done is create a lot of special purpose topics without properly setting up the upper level groups needed to allow general purpose topics as well-- if we do that, the need for special purpose topics goes away unless they generate enough interest to warrant splitting them off. I'm looking at the name space now, I think I'll have a first draft of a suggested re-arrangment for public comment soon (It isn't an easy task, unfortunately). The total number of groups will probably stay about the same, or maybe even grow a bit-- whatever it takes to rebuild the name space coherently. > The presence of a news group encourages traffic. For instance net.music is >subdivided, the latest sub group being net.music.synth. It seems that there are >a lot more articles on synthesis now than before the group was created, and it >grew all at once to a well used group. Some rarely used groups such as .poetry >have justification in that although few post to it, it is also true that few >are poets or are brave enough to post their own poems. The fact that net.music.synth has a lot of volume shows that there is a lot of interest for it, so creating it was a good idea. If the volume goes away with time (as, for example, it has with net.games.go) it should also go away again, and the discussion merged back with the parent group. We currently don't have a good policy for doing that, but I think it is time for doing so. A big problem with that is, of course, the fact that most groups don't have appropriate parents to merge into, and probably should. chuq -- Chuq Von Rospach, National Semiconductor {cbosgd,fortune,hplabs,ihnp4,seismo}!nsc!chuqui nsc!chuqui@decwrl.ARPA Be seeing you!