jim@eda.com (Jim Budler) (11/09/89)
coolidge@brutus.cs.uiuc.edu (John Coolidge) writes: } davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) writes: } > I think it is safe to say that B news is a more mature package than C } >news. It is faster. In my opinion it is harder to install on things } >other than BSD, because it has not been worked on by as many sites. } >That's not a complaint, just a statement of fact, it's new. } This is true, as long as the 'It' in the second and following } sentances means 'C News'. The first time I read it, it parsed as } 'B News' and didn't make much sense. C news is faster, is a bit } less well worked on, and is much newer. I have to agree with this, completely. The parsing of the referenced article almost led me to hitting the f (flame) key. Suffice it to say: I sit at a Sun386i, it has our modem to the outside world, it runs our news software, it has our news disk. I have a constant perfmeter display on my screen. Under B-news the end of a transmission from my newsfeed, or unbatch of news (I tried both methods) my perfmeter hit 100% and stayed there. Under C-news it hits 100% for startup, then drops to 50% and stays there. In addition, and this is subjective, I feel the total process time is less under C-news, i.e. I spend less time at 50% CPU than I used to spend at 100%. This last could be totally subjective. I conclude C-news is more efficient. None of my newsreader software even noticed the change. We use rrn and nn here. } > The V3 of B news will be out shortly, and it offers the same } >performance gains as C news, although obviously not in the same places. Shortly counts for nothing. I've been using C-news for 2 months. } >Because it was developed on a SysV machine it seems a lot more solid } >there than C news. That's nice, no oppinion. [...] } C News has worked very well for us. I've done a bunch of tweaking } and patching, but most of that is due to really high-performance } nntp concerns. As a fully-functional news setup C worked just fine } 'out of the box'. Not ever having run B news, we never had any } trouble with old B traps such as using inews -C to create groups, } or thinking things had to be done the B news way. I'm happy with } C news and would recommend it for new sites (and I've even run } it on a SYSV box with no problems, once upon a time...). Having run B-news I think the new addgroup and delgroup commands are *wonderfull*. Your average news admin can easily add a new local group without sweating over a manual for a half hour. *Everything* you do with addgroup or delgroup is by definition *local*. If you are a wizard and want to do something more you have to build the RFC compliant file containing the appropriate headers and information and feed it to 'inews -h'. Elegant: regular stuff is provided for, wizard stuff requires wizard preparation. Unlike B-news where all operations were wizard operations. } --John Me: My Opinion: jim -- Jim Budler jim@eda.com ...!{decwrl,uunet}!eda!jim compuserve: 72415,1200 applelink: D4619 voice: +1 408 986-9585 fax: +1 408 748-1032
peter@ficc.uu.net (Peter da Silva) (11/10/89)
In article <1989Nov9.102613.16627@eda.com> jim@eda.com (Jim Budler) writes: > *Everything* you do with > addgroup or delgroup is by definition *local*. If you are a wizard > and want to do something more you have to build the RFC compliant > file containing the appropriate headers and information and feed it > to 'inews -h'. Bletch. > Elegant: regular stuff is provided for, wizard stuff > requires wizard preparation. Unlike B-news where all operations > were wizard operations. Oh, fiddlesticks. What's so wizard about "inews [-d dist] -C newsgroup"? We create organisational-level groups all the time, with no wizard preparations. I'll eventually go to cnews, after it's gone a while without patches. But I'll have to say it doesn't look very convenient at all for anything but a plain-jane single-machine setup. -- `-_-' Peter da Silva <peter@ficc.uu.net> <peter@sugar.hackercorp.com>. 'U` -------------- +1 713 274 5180. "*Real* wizards don't whine about how they paid their dues" -- Quentin Johnson quent@atanasoff.cs.iastate.edu
henry@utzoo.uucp (Henry Spencer) (11/12/89)
In article <6906@ficc.uu.net> peter@ficc.uu.net (Peter da Silva) writes: >I'll eventually go to cnews, after it's gone a while without patches. But >I'll have to say it doesn't look very convenient at all for anything but >a plain-jane single-machine setup. Can you elaborate on this? We've got pretty good support for NFSed multi- machine setups, since one of our development systems (Geoff's) is such. -- A bit of tolerance is worth a | Henry Spencer at U of Toronto Zoology megabyte of flaming. | uunet!attcan!utzoo!henry henry@zoo.toronto.edu
ambar@bloom-beacon.mit.edu (Jean Marie Diaz) (11/13/89)
From: henry@utzoo.uucp (Henry Spencer) Date: 12 Nov 89 01:05:39 GMT In article <6906@ficc.uu.net> peter@ficc.uu.net (Peter da Silva) writes: >I'll eventually go to cnews, after it's gone a while without patches. But >I'll have to say it doesn't look very convenient at all for anything but >a plain-jane single-machine setup. Can you elaborate on this? We've got pretty good support for NFSed multi- machine setups, since one of our development systems (Geoff's) is such. It's also pretty easy to run a NNTP-based system, iff you understand (ie, are already running) NNTP. (Neither C News nor NNTP is documented with the novice news admin in mind.) In another posting, Rick Adams points out that the availability of O'Reilly's _Mananging UUCP and Usenet_ is a point in B News' favor. I should point out that the eighth edition (which should be available before the new year) will cover both B News 2.11 and C News. (Claimer: of course I'm biased, I did the rewrite :-) AMBAR ambar@bloom-beacon.mit.edu {mit-eddie,uunet}!bloom-beacon!ambar
peter@ficc.uu.net (Peter da Silva) (11/13/89)
How about uucp-ed multi-machine setups hiding behind a single name, with a single password file and user base? -- `-_-' Peter da Silva <peter@ficc.uu.net> <peter@sugar.hackercorp.com>. 'U` -------------- +1 713 274 5180. "*Real* wizards don't whine about how they paid their dues" -- Quentin Johnson quent@atanasoff.cs.iastate.edu
henry@utzoo.uucp (Henry Spencer) (11/16/89)
In article <6936@ficc.uu.net> peter@ficc.uu.net (Peter da Silva) writes: >How about uucp-ed multi-machine setups hiding behind a single name, with >a single password file and user base? Assuming that the machine don't share filesystems, the only real issue here for a news system is the need to hide the internal structure from the outside world. This will mean some tinkering with Path headers and such on outbound news. It should not be very hard to fit such a thing in -- it would mean using a smarter variant of the "batcher" program for external feeds -- but we don't have it off the shelf. Neither of us is familiar enough with such systems to be sure exactly what's wanted. -- A bit of tolerance is worth a | Henry Spencer at U of Toronto Zoology megabyte of flaming. | uunet!attcan!utzoo!henry henry@zoo.toronto.edu