Xref: utzoo unix-pc.general:6564 comp.sys.att:10964
Path: utzoo!utgpu!watserv1!watmath!uunet!shelby!apple!vsi1!zorch!hico2!kak
From: kak@hico2.UUCP (Kris A. Kugel)
Newsgroups: unix-pc.general,comp.sys.att
Subject: Re: CALL FOR DISCUSSION: comp.sys.3b1
Summary: We're weird here too.
	 (followups to late and missing original articles)
Message-ID: <375@hico2.UUCP>
Date: 26 Nov 90 17:09:06 GMT
References: <1990Nov19.040128.29048@bagend.uucp> <1085@galaxia.Newport.RI.US>
Followup-To: unix-pc.general,comp.sys.att
Organization: High Country Software
Lines: 55

In article <1085@galaxia.Newport.RI.US>, dave@galaxia.Newport.RI.US (David H. Brierley) writes:
> I think one of the things that this discussion has made clear to me is that
> sites that thought they were well connected in terms of unix-pc.* are in
> fact not well connected.  I have eight separate news feeds sending me the
> unix-pc.* groups (and conversely, sending my stuff back out) and yet I have
> seen several replies that I never saw the base article to.  Also, I posted
> a message saying that I was taking an informal survey which would be followed
> by a formal call for discussion if the survey seemed to warrant it and yet
> two separate people have posted similar messages (one survey and this CFD).
> I know that my message made it out because I got almost two dozen replies
> within a few days, almost all in favor of the proposed new group.

Timing on the net is "weird and wonderful."  For example,
I feed unix-pc to 6 different sites. 

Only one of those feeds is a "pure" feed, sending everything I get.
All the rest are filtered somehow.  One other site is ihave/sendme,
(so they SHOULD get everything eventually), and several others are
filtered to remove articles they should have gotten otherwise.
For example, I don't forward those unix-pc articles to zorch that
have passed through uunet, att, or rutgers, figuring that zorch
can get these articles through uunet and the internet easier 
(and cheaper) than me.  I also limit the number of machines
that an article can pass through before me and still get to zorch,
figuring that zorch will get articles from further away through
other means.  I try to get my and my near neighbors' articles out
to zorch, it's worth the expense to increase reliability.
Almost all of the articles I send out are batched/compressed, and
the times they get sent out cause oddness in the news flow too.

Now if one of my neighbors replys to an article from far away,
I'll send the followup to zorch but not the original.
So the original might get to zorch's neighbors later than the
reply or not at all.

> Second, comp.sources.3b1
> (moderated) for postings of source programs specific to the 3b1 and for
> patches to those programs. 

One problem with this is that moderated groups depend on the mail path
to the moderator being open.  News is more reliable than mail.
Posting directly is also much faster distribution, and less of a
single-person bottleneck.

I wouldn't mind a moderated archive group, distributed only between
archive sites.  (alt.3b1.sources?)

> Oh yeah, once the new groups are created (assuming that they are), the
> old unix-pc.* groups should be disbanded.

Yes, one or the other but not BOTH.

                               Kris A. Kugel
                             ( 908 ) 842-2707
   { uunet | rutgers }!{ tsdiag | westmark }!hico2!kak
                                att!westmark!hico2!kak
                         {daver,ditka,zorch}!hico2!kak
                         internet: kak@hico2.westmark.com