andrews@ubc-cs.UUCP (Jamie Andrews) (06/17/86)
I just posted to net.sources a shar package for the Junker program,
described earlier on net.news.group. If you haven't heard of this
before, it's a program which "junks" certain specified newsgroups
by slicing out the middles of large articles until the total size
of articles is below a certain specified limit. Used judiciously it
has the effect of reducing total traffic across communication lines
for the newsgroups.
If you start using it, please send me a note at one of the
addresses below. I can understand if people don't want to jump into it
without some net-wide use of it on certain newsgroups; I would appreciate
it if a group of backbone or near-backbone people would start using it
on some newsgroups no one cares about, like net.flame or net.bizarre (if
they're still alive somewhere). I can even imagine people increasing
the connectivity of net.flame as a torture test for junker. Then junker
can spread to more sites and more newsgroups.
For exactly $0.00 (+ transmission costs of course!) you get in the
shar package:
- the full public-domain listing of junker
- a man entry
- a makefile
- information on installation (minimal, because it was designed with
ease of use in mind)
- information for net users on junker
If the use of junker becomes widespread, people will start seeing
parts of articles cut out. It's so hard to predict what the effect of
this will be; what I would want it to be, and what I'm fairly confident
it will be, is this. After the initial backlash against it, people on
the affected newsgroups (probably the "soapbox" newsgroups) will start
understanding that junker is better than newsgroup cuts. Junker cuts
out stuff regardless of whether it's signal or noise, and the newsgroup
readers will realize this. Each affected newsgroup will become a
self-policing community which will work to cut down the volume of noise,
in order that signal not be cut out.
At least that's the theory. Even if it doesn't happen, volume will
be reduced by junker. And since junker is designed to be painless to
install, modify, and get rid of, if it all doesn't work it'll be easy to
chuck it away.
Until the end of the summer, and probably until the end of 1986, I
will be supporting junker. Please send comments, suggestions, and
especially bug reports to me. Followups will go by default to
net.news.adm.
--Jamie Andrews.
...!{ihnp4!alberta | seismo}!ubc-vision!ubc-cs!andrews
or andrews@cs.ubc.cdn
"That was my Lo... and these are my lilies"wmartin@brl-smoke.ARPA (Will Martin ) (06/20/86)
Does no one out there realize what a terribly bad idea this "junker" concept is? It is deliberately introducing transmission bugs into an already-fragile and problem-ridden system! There is absolutely no basic difference between what "junker" is supposed to do (chop the middle out of otherwise good [e.g, complete] articles) and what the line-eater bug did [or does] (chop the front off otherwise complete articles). Not only is this an awful idea to deliberately adopt, but the net effect is an increase in net-wide costs -- people are now going to be paying to transmit garbled and incomplete articles, the subsequent reports of the garbling, the retransmission attempts as people try to get their articles sent in one piece, extra header overhead as people split long articles into several separate parts, and so on. When I first read the tentative proposal, it was obvious to me that it was a bad idea. It cannot possibly do any good. Why do not the rest of you see this clear fact? I cannot believe that this bad idea has progressed so far! Will
grr@cbmvax.cbm.UUCP (George Robbins) (06/20/86)
In article <282@ubc-cs.UUCP> andrews@ubc-cs.UUCP (Jamie Andrews) writes: > > I just posted to net.sources a shar package for the Junker program, >described earlier on net.news.group. If you haven't heard of this >before, it's a program which "junks" certain specified newsgroups >by slicing out the middles of large articles until the total size >of articles is below a certain specified limit. Used judiciously it >has the effect of reducing total traffic across communication lines >for the newsgroups. > I think that anybody who feeds sites outside of their organization had better consider very carefully whether or not they should use this program. It's OK to limit your own internal costs, but it's far too easy for these mutilated articles to get back out on the net and interfere with the transmission of the intact article. This has about all the appeal of another notes blowout or a resurgence of the line eater bug. Let's try to improve the net through some kind of generally acceptable administrative changes, not through careless vigilante sabotage. -- George Robbins - now working with, uucp: {ihnp4|seismo|caip}!cbmvax!grr but no way officially representing arpa: cbmvax!grr@seismo.css.GOV Commodore, Engineering Department fone: 215-431-9255 (only by moonlite)
avolio@decuac.DEC.COM (Frederick M. Avolio) (06/20/86)
In article <282@ubc-cs.UUCP> andrews@ubc-cs.UUCP (Jamie Andrews) writes: > I just posted to net.sources a shar package for the Junker program, >described earlier on net.news.group. If you haven't heard of this >before, it's a program which "junks" certain specified newsgroups >by slicing out the middles of large articles until the total size >of articles is below a certain specified limit. Actually, I originally thought that this idea was kind of a joke and never gave it much notice. Reminds me of something I put in a message-of-the-day file when I was a system manager a number of years ago. I was something about trying to free-up some disk space by utilizing the Hoffman Prime Number Block Algorithm (or some such nonsense) which would free up all data blocks in files which had a prime number, thus freeing up some much needed space. One person called to ask for an explanation. It was, of course, April 1st. -- Fred @ DEC Ultrix Applications Center INET: avolio@decuac.DEC.COM * Fight the Fight * UUCP: {decvax,seismo,cbosgd}!decuac!avolio * Rescue the Unborn *
henry@utzoo.UUCP (Henry Spencer) (06/22/86)
As a matter of policy, utzoo (a backbone site) will not run the junker or
any similar program. We are concerned about volume: we have cut some of
the "noisier" newsgroups, we may cut more, and we strongly support moderated
groups. But we don't think this particular idea will work well enough to
be worth the damage it will cause.
--
Usenet(n): AT&T scheme to earn
revenue from otherwise-unused Henry Spencer @ U of Toronto Zoology
late-night phone capacity. {allegra,ihnp4,decvax,pyramid}!utzoo!henryhenry@utzoo.UUCP (Henry Spencer) (06/26/86)
A further refinement of our policy on the junker: We will not run the
junker or any similar program. If any of the sites we feed starts running
it, they will have to find a new feed. We will encourage the sites we
feed to adopt the same policy.
--
Usenet(n): AT&T scheme to earn
revenue from otherwise-unused Henry Spencer @ U of Toronto Zoology
late-night phone capacity. {allegra,ihnp4,decvax,pyramid}!utzoo!henryspaf@gatech.CSNET (Gene Spafford) (06/27/86)
In article <282@ubc-cs.UUCP> andrews@ubc-cs.UUCP (Jamie Andrews) writes: > > [description of the junker] > > If you start using it, please send me a note at one of the >addresses below. I can understand if people don't want to jump into it >without some net-wide use of it on certain newsgroups; I would appreciate >it if a group of backbone or near-backbone people would start using it When Jamie first posted his proposal about writing the "junker," I didn't take it too seriously -- I thought the feedback others provided was sufficient to point out that the concept was undesireable. I really, really appreciate the fact that he has put some thought into the problem and even went to the effort of producing some code to implement his fix, but his solution is much worse than the problem itself. In my opinion, mucking about with the contents of news is outright vandalism, especially since the current software structure means that reception of a mangled copy of the article will prevent an intact version from arriving via another route. To see if I was the only one objecting to the junker, I polled the other backbone admins and the moderators by mail. I have received 22 responses so far, all but two from backbone admins. Every single one is vehemently against the junker. Many suggested attempts at reprisals against any site "junking" articles in "mod" groups or net.sources. Most (19) felt that junking *any* newsgroup was offensive. I suspect that this view is not in the minority. I asked them what their comments were on someone else's (Charlie Wingate) posted comment that he wouldn't even forward mail to sites running the junker. Many supported that or a similar stance. Some of their edited comments are included below. I have omitted names and other identifying info. Summary: we appreciate efforts to help reduce traffic, but mangling and vandalizing news is not one of those methods. Sites are obviously free to run any software they choose, but running the "junker" or any other software causing widespread damage to the integrity of the Usenet does so at their own risk and with active opposition from the majority (if not the totality) of the backbone and many associated sites. Comments: ---- ...maybe [I'd] break into his system and run junker on /etc/passwd ---- Anti junker. All the way. Nuke it til it glows and shoot it in the dark. ---- I wouldn't run it here, and I'd discourage others from running it, but not having the hooks or time to beat on them, I personally wouldn't go to any superhuman efforts to stop others from running it. ---- If I found a neighbor newsfeed junking articles, I would turn off the link. We have gone to so much effort to insure that news articles are transmitted intact. The "junker" approach seems to be completely against the spirit of: the net, linking of sites, accurate data transmission, mutual cooperation. etc. "Junker" is little-boy spitefulness expressed in computer code. I prefer to work in a different (adult?) mode. ---- I will not, under any circumstances, run 'junker'. ---- I think 'junker' is bad for the net. It is a simple-minded attempt to solve a complex problem, rather than using some brainpower and coming up with an intelligent solution. It also, in my opinion, will not work. It will never be installed at [my site]. Not for technical, non-technical, talk, soapbox or any other newsgroup. It will make USENET into a joke. I prefer rmgroup to junker. It doesn't make any sense. Kill it before it multiplies. Need I say more? ---- traffic, but it is quite another to alter it and send it on down the line--is this significantly different from adding nasty comments into a message? And, wasn't there once a discussion about being a common carrier? That is, I was under the impression that [my site] wouldn't be taken to court if we pass traffic from other places that might be considered to be illegal in nature. Wouldn't we lose this standing if we started applying *any* editorial control beyond the decision to carry or not to carry? ---- I ignored the [pejorative deleted], not believing he could be serious. Woe to me for again overestimating the idiocy of the network... ... it will just cause a new set of "such and such an article was truncated, please repost" followed by repostings. The people it is designed to go after will quickly learn to get around it. pheh. I'd be tempted to hack my Path to make sure they never SAW an article of mine, and perhaps hack my mailer to not only not forward to them, but to eat anything enroute to them regardless of where it went next. The thing is nasty and insidious. Take a look at how the whitespace bug in 2.10.2 affected the net, and think about how a conscious and intentional bug would affect things. In many respects, the two are the very same. ---- If they were sites I fed, I would stop feeding them. I might be tempted to drop them from my L.sys..... I guess the problems with the junker have already been publicized: the fact that it's too easy to work around, that the offender (or their SA) is never told, etc. I particularly dislike the way it lets sites "pretend" to offer full service, while in fact they aren't. At least now, if a site doesn't care net.rec.drugs, you know you're not getting the whole thing. ---- Discarding articles might be acceptable in some cases, but mutilating them is wrong both morally and as a technical solution. No news connections for sites that mutilate articles. ---- Junk Junker. ---- I would cut the UUCP connection entirely. I would call on the management of the site [running junker] to persuade the offending party to cease and desist. If this was ineffective, I'd call the management of sites that they fed and ask that they find an alternate feed. I'd call the management of sites that feed the offending site and ask that his feed be terminated. I'd also give the offending party's name to every headhunter I knew... :-) ---- -- Gene Spafford Software Engineering Research Center (SERC), Georgia Tech, Atlanta GA 30332 CSNet: Spaf @ GATech ARPA: Spaf@Gatech.GATECH.EDU uucp: ...!{akgua,decvax,hplabs,ihnp4,linus,seismo,ulysses}!gatech!spaf
wsr@lmi-angel.UUCP (Wolfgang Rupprecht) (07/03/86)
Clearly junker should be (repeatedly) applied to its own sources. Tis
only fitting.
Why not put a quota on newsgroups, with those extra articles put in
*tommorrows* send queue? Since the number of articles that people
write increases with the number that they read, this would have a
similar effect to lowering the moderation rods (in a nuclear reactor).
Is this just as crazy?
--
Wolfgang Rupprecht {harvard|decvax!cca|mit-eddie}!lmi-angel!wsr