bryce@eris.berkeley.edu (Bryce Nesbitt) (03/03/88)
In article <> farren@gethen.UUCP (Michael J. Farren) writes: > >I am going to urge a NO vote on comp.binaries.hypercard. My reasons? >Basically, two. First, the volume... >...Second, the limited applicability of this... > >If you take this as a condemnation of binary groups in general, you're >probably right. It would be very intersting to see how many times computer-specific binary messages are actually downloaded by someone.* I strongly suspect is is often a much smaller number than the magic "100 message - mail vs. net breakeven point". Perhaphs what is needed is a demand based system.** A reader sees a *full* description of the binary, and presses a key that means "I want a copy". This message is stored by rn, and transmitted upstream next time news is processed.*** The delay is unimportant for most binaries, (especially long binaries).**** * Yes, I am aware how impossible this would be under the current system. ** Yes I know that it would be much work to add. Also I know, but do not like, the fact that not all sites run the same software, and that may sites run old software. (-: I'll bet more would upgrade if the new system was made incompatible :-). *** A backchannel would be great for reader-based "ratings" on the quality of articles. **** Yes, there is justification for this. The only urgent binaries are likely to be patches. Patches are usually small. |\_/| . ACK!, NAK!, EOT!, SOH! {O_o} . Bryce Nesbitt (") BIX: mleeds (temporarily) U USENET: bryce@eris.berkeley.EDU -or- ucbvax!eris!bryce
cyrill@scicom.AlphaCDC.COM (Cyro Lord) (05/06/89)
The one feature that I would like to see in the news-software
(version 3?) would be the ability to use the 'active' news files
to dump newsgroups into the VOID! (/dev/null) That way, the feed
never needs to be modified or updated for partial feeds or for full
feed sites, they could get rid of newsgroups they don't feed or
need. A way this could work is the newsgroup would have a '!' either
in front of it, in place of the 0's or replacing the 'y' or 'm'.
Of course, to turn it on again, one need only change back again.
What do you think??
--
Cyro Lord Alpha Comm. Dev. Corp. - DOMAIN cyrill@scicom.alphacdc.com
UUCP {ncar,nbires,boulder,isis}!scicom!cyrill
Corn can't expect Justice from a court of chickens. (African Proverb)tale@pawl.rpi.edu (David C Lawrence) (10/04/89)
In <1989Sep30.050659.21928@alembic.acs.com> csu@alembic.acs.com (Dave Mack):
DM> Wrong, at least in news2.11 and C news. If an article is received with
DM> no valid news groups, it goes into junk and isn't broadcast to neighboring
DM> sites, regardless of the Distribution field. NNTP may behave differently.
NNTP doesn't behave differently because it doesn't behave on that
level at all. NNTP is just a transport mechanism; it interacts with
the existing "News Overlord Software" which makes the real decisions
about where to put articles and to whom to forward them.
Just trying to clear up a misunderstanding before it spreads too far.
Follow-ups to news.software.b which has become a general news.software
group. Perhaps it is time to really change the name.
Dave
--
(setq mail '("tale@pawl.rpi.edu" "tale@itsgw.rpi.edu" "tale@rpitsmts.bitnet"))pst@anise.acc.com (Paul Traina) (10/04/89)
In <1989Sep30.050659.21928@alembic.acs.com> csu@alembic.acs.com (Dave Mack): DM> Wrong, at least in news2.11 and C news. If an article is received with DM> no valid news groups, it goes into junk and isn't broadcast to neighboring DM> sites, regardless of the Distribution field. NNTP may behave differently. Sorry to correct you Dave, but in at least some instances of C news, the article is still broadcast. I noticed this with the newgroup control message (perhaps a special case?) for comp.dsp. -- Insanity is the exception in individuals. In groups, parties, people, and times, it is the rule. -- Nietzche
mack@inco.UUCP (Dave Mack) (10/05/89)
In article <1989Oct4.153218.14076@ginger.acc.com> pst@anise.acc.com (Paul Traina) writes: >In <1989Sep30.050659.21928@alembic.acs.com> csu@alembic.acs.com (Dave Mack): >DM> Wrong, at least in news2.11 and C news. If an article is received with >DM> no valid news groups, it goes into junk and isn't broadcast to neighboring >DM> sites, regardless of the Distribution field. NNTP may behave differently. > >Sorry to correct you Dave, but in at least some instances of C news, >the article is still broadcast. I noticed this with the newgroup control >message (perhaps a special case?) for comp.dsp. My apologies - yes, it appears that non-local control messages always propagate. There may be a couple of other exceptions as well, but in general, an article which has no valid newsgroups will be junked and not be sent to other sites. -- Dave Mack McDonnell Douglas Electronic Systems uunet!inco!mack, inco%mack@uunet.uu.net (703)883-3911 All opinions expressed are my own and do not reflect those of MDESC. Ever.
geoff@utstat.uucp (Geoff Collyer) (10/05/89)
Dave Mack: > Wrong, at least in news2.11 and C news. If an article is received with > no valid news groups, it goes into junk and isn't broadcast to neighboring > sites, regardless of the Distribution field. This is false in the case of C news. Articles may be filed in the control or junk pseudo-groups, but that has no effect on their distribution (see relaynews(8)). Your active file may be missing some groups, while your neighbours have those groups in their active files. If you really want to drop articles in certain groups, and ignoring cross-posting for now, you must either write "!unwanted.group" in your sys file subscription list, or you must change the flags field of "unwanted.group" to "x" (see news(5)). -- Geoff Collyer utzoo!utstat!geoff, geoff@utstat.toronto.edu
csu@alembic.acs.com (Dave Mack) (10/08/89)
In article <1989Oct4.234355.3680@utstat.uucp> geoff@utstat.uucp (Geoff Collyer) writes: >Dave Mack: >> Wrong, at least in news2.11 and C news. If an article is received with >> no valid news groups, it goes into junk and isn't broadcast to neighboring >> sites, regardless of the Distribution field. > >This is false in the case of C news. Articles may be filed in the >control or junk pseudo-groups, but that has no effect on their >distribution (see relaynews(8)). Your active file may be missing some >groups, while your neighbours have those groups in their active files. >If you really want to drop articles in certain groups, and ignoring >cross-posting for now, you must either write "!unwanted.group" in your >sys file subscription list, or you must change the flags field of >"unwanted.group" to "x" (see news(5)). Urk! Sorry, I botched the sense of the test of okrefusal. I thought the ST_DROPPED bit was getting set. OK, so news2.11 junks articles with no valid newsgroups, C news passes them on. I can see arguments in favor of each strategy. With news2.11, it's possible to rmgroup a group and actually have it gradually disappear; with Cnews, everything still circulates unless you take steps to squash it in your sys file. On the other hand, with Cnews, a site has more trouble blocking transmission of a legit group that it doesn't like. Since I'm not overly concerned with newsgroup deletion but am concerned with the maximum propagation of "real" newsgroups, I suggest that the way C News does it is correct. Comments? Dave Mack
greyham@hades.OZ (Greyham Stoney) (10/10/89)
in article <5654@inco.UUCP>, mack@inco.UUCP (Dave Mack) says: > ......... There may be a couple of other exceptions as well, but in > general, an article which has no valid newsgroups will be junked and > not be sent to other sites. Oh.... yeah?. We run Bnews (I think) and man inews says (talking about inews -p): When inews receives an article this way, it will check the history file to make sure that the article is not already present, and it will make certain consistency checks to make sure that the newsgroup names are legal and that the sys file permits the article to be installed on the local machine. Once the article passes those checks, it is installed in the appropriate directory on the local machine. If the article fails those checks, it is installed in news- group junk on the local machine. In any event, inews will ^^^^^^^^^^^^^^^^^^^^^^^^ then transmit the article to all systems that match in the ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ sys file and are not mentioned in the Path: field of the ^^^^^^^^ just-posted message. The details of this transmission are determined by the contents of the sys file. Hmmmm? Greyham. -- /* Greyham Stoney: Australia: (02) 428 6476 * * greyham@hades.oz - Ausonics Pty Ltd, Lane Cove, Sydney, Oz. * * TDMP/IP: Telepathic Direct Marketing Personel Interface Protocol */
mack@inco.UUCP (Dave Mack) (10/13/89)
In article <457@hades.OZ> greyham@hades.OZ (Greyham Stoney) writes: >in article <5654@inco.UUCP>, mack@inco.UUCP (Dave Mack) says: >> ......... There may be a couple of other exceptions as well, but in >> general, an article which has no valid newsgroups will be junked and >> not be sent to other sites. > >Oh.... yeah?. We run Bnews (I think) and man inews says (talking about >inews -p): > [quoting inews man page:] > If the article fails those checks, it is installed in news- > group junk on the local machine. In any event, inews will > ^^^^^^^^^^^^^^^^^^^^^^^^ > then transmit the article to all systems that match in the > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ > sys file and are not mentioned in the Path: field of the > ^^^^^^^^ > just-posted message. The details of this transmission are > determined by the contents of the sys file. > > Hmmmm? > Greyham. Sorry, I didn't look at the man page. I read the code, which says, in part (inews.c/patchlevel17:insert()): ... is_invalid = ngfcheck(username, mode == PROC, header.approved[0] != '\0'); ... if (is_invalid) { logerr("%s: No valid newsgroups found, moved to junk", header.nbuf); if (localize("junk")) savehist(histline); exitcode = 1; goto writeout; } ... writeout: ... if(exitcode == 0) { ... broadcast(mode==PROC); ... } xxit((mode == PROC && filename[0] == '\0') ? 0 : (exitcode < 0 ? 0 : exitcode)); } All of this is ifndef NFSCLIENT, of course. -- Dave Mack McDonnell Douglas Electronic Systems uunet!inco!mack, inco%mack@uunet.uu.net (703)883-3911 All opinions expressed are my own and do not reflect those of MDESC. Ever.