[can.usrgroup] see news?

evan@telly.on.ca (Evan Leibovitch) (04/06/89)

What IS the status of Cnews? There was a 'full release' version due at
Usenix but I heard it was delayed. Is it ready yet?

If not, what's the latest release level, and what are the holes which
the 'release' version is expected to fill?

I also notice that TMN news is expected out soon. Does anyone anticipate
using the inews/expire mechanisms from Cnews and readers/posters from
TMN, or is it better to stick with one throughout?

Thanks.

geoff@utstat.uucp (Geoff Collyer) (04/07/89)

In article <8904052337.AA23221@telly.on.ca> evan@telly.on.ca (Evan Leibovitch) writes:
>What IS the status of Cnews? There was a 'full release' version due at
>Usenix but I heard it was delayed. Is it ready yet?

C news is mostly in private beta test; we are waiting for Henry's expire
before the beta test kit will be complete.  The delay has been due to
Henry and I being sick, an emergency in my family that kept me out of
town for most of a month, and my inability to put fire in Henry's belly
earlier :-).

>If not, what's the latest release level, and what are the holes which
>the 'release' version is expected to fill?

The current public release is *still* C alpha; the production release
fixes a long list of omissions and some bugs.  In particular, full
disks, control messages and moderated groups are treated correctly,
there are better mechanisms for coping with problems in the groups
being sent by your feed, the batcher is simpler than in the alpha
release, the input subsystem copes with #! c7unbatch (or whatever),
expire will be able to retain history longer than the articles,
essentially all the software will have no fixed-size buffers to be
overflowed, there is support for i-have/send-me (batched only) and NNTP
(batched upon receipt), and there is finally some installation
documentation.  (I may have missed something, the alpha release was a
long time ago, as if you needed reminding.)  It's still fast, though
most of our energy since the alpha release has been on robustness and
completeness.   Someone locally has even ported the alpha release to
MS-DOS(!); no, I don't know how he did it.

One thing that may not be finished in time for the production release is
a faster inews (as opposed to rnews).  It's not much of a problem for
posting a single article (even Henry says so :-), but for gatewaying
mailing lists into newsgroups, something faster would help machines such
as jarvis.csri.toronto.edu (nee utjarvis) or ucbvax or uunet.
Unfortunately, inews is also where most of the site-specific
customisation happens, so it's a difficult trade-off.  However the
screams from CSRI are a little unsettling, and some of those folks run
one of the major U of T mail relays (neat.ai.toronto.edu, utai to most
of you) and they keep reminding me about the problem at dinner :-), so
we will likely provide at least a minimal hook for a faster inews.

>I also notice that TMN news is expected out soon. Does anyone anticipate
>using the inews/expire mechanisms from Cnews and readers/posters from
>TMN, or is it better to stick with one throughout?

I imagine that Henry and I will at least look at what the competition
has been up to, in part because we supply only a trivial readnews, but
it will take something major (such as power, compactness, simplicity and
speed) to shift us from using rn, which is certainly powerful.

P.S. Henry and I have been considering the invitation to speak at UU on
the insides of C news when this is all over.  Talking about your work is
much easier than doing it.
-- 
Geoff Collyer	utzoo!utstat!geoff, geoff@utstat.toronto.edu
It's all Henry's fault. (TM)