amanda@mermaid.intercon.com (Amanda Walker) (08/06/90)
In article <1990Aug04.171540.29439@looking.on.ca>, brad@looking.on.ca (Brad Templeton) writes: > The problem is that the net is not a NEW and powerful idea. I've been > on it just shy of a decade myself. In the world of computer networking, > a decade is almost forever. Indeed. This net is new to a great many people, by virtue of its growth, but that very size has given it a huge supply of inertia. I remember the A News to B News changeover in '81-'82, and it was a royal pain, even though Usenet was *much* smaller then, and was still thought of as an experiment. With the state of the net as it is today, I don't think that any major changes are really possible, except by building something better that renders Usenet obsolete. Look at C news--it's ingratiatingly compatible with B news, and the only people that seem to be picking up on it are us performance weenies out here :-). It's only gaining widespread acceptance very slowly, and that only because it's a fairly "invisible" upgrade. I don't think that any kind of format change, even one as small as Brad's Favorite Headers (tm), is going to catch on, however useful it may be. Too many people see Usenet as it is to be "useful enough," and don't see any reason to change until there is already something better in place and working. It's annoying in some respects, but I haven't seen any contrary evidence so far. Anyone have "wish lists" for a kinder, gentler Usenet, assuming backward compatibility isn't an issue? -- Amanda Walker <amanda@intercon.com> InterCon Systems Corporation
henry@zoo.toronto.edu (Henry Spencer) (08/08/90)
In article <26BCDA30.21C@intercon.com> amanda@mermaid.intercon.com (Amanda Walker) writes: >... Look at C news--it's ingratiatingly compatible >with B news, and the only people that seem to be picking up on it are us >performance weenies out here :-). It's only gaining widespread acceptance >very slowly, and that only because it's a fairly "invisible" upgrade... Actually C News isn't doing too badly; based on analysis of message-IDs in our history file, C News is running at 600+ sites, about 12% of the active subset of Usenet. This method inherently can't measure the passive subset -- the sites that rarely or never post anything -- but we know C News gets substantial use among them too. (The rest of the active subset is currently split about 50-50 between B News sites and sites generating weird message-IDs that aren't obviously either B or C. There is no single dominant form of weirdness. :-)) Both the count and the percentage for C News are growing steadily. However... Amanda is right, steady it may be, but it's also slow. It would probably be somewhat faster if we were sysadmin-compatible with B News, because learning new procedures increases the conversion effort. This bodes ill for any change that involves incompatibility of a more serious kind. Yes, a lot of this is sheer inertia, as witness the sites that are still running B2.6, but each obstacle to conversion greatly magnifies that inertia. A seriously-incompatible news system will have little chance unless it offers really major benefits to make up for the hassle. -- The 486 is to a modern CPU as a Jules | Henry Spencer at U of Toronto Zoology Verne reprint is to a modern SF novel. | henry@zoo.toronto.edu utzoo!henry
emv@math.lsa.umich.edu (Edward Vielmetti) (08/08/90)
gee henry, there are a lot of known bugs in B News that cause it to dump core or something like that. just flood the net regularly and predictably with these articles and it should wipe out the worst of the sites pretty quick. Just make sure C News passes the torture test.... --Ed Edward Vielmetti, U of Michigan math dept <emv@math.lsa.umich.edu>
fitz@wang.com (Tom Fitzgerald) (08/09/90)
henry@zoo.toronto.edu (Henry Spencer) writes: > based on analysis of message-IDs > in our history file, C News is running at 600+ sites, about 12% of the > active subset of Usenet. This is probably a serious underestimate, given the number of sites who: - Are running mail-to-news gateways and use the mailmessage's Message-ID in the article. - Don't like C-news's super-jumbo message IDs, and have modified inews to generate something more concise (like us). Also, don't some NNTP posting agents generate their own message-IDs in the posting program, rather then on the NNTP server? This would hide another population of C-news sites. --- Tom Fitzgerald Wang Labs fitz@wang.com 1-508-967-5278 Lowell MA, USA ...!uunet!wang!fitz
henry@zoo.toronto.edu (Henry Spencer) (08/09/90)
In article <ar37h5.c50@wang.com> fitz@wang.com (Tom Fitzgerald) writes: >This is probably a serious underestimate... >... mail-to-news gateways ... >...[avoidance of] super-jumbo message IDs... >... NNTP posting agents ... Yes, it's very difficult to assess these things. However, message-ID analysis has the enormous virtue that it's quite cheap, unlike the fairly drastic step of sending out a "version" control message. Message-IDs only give us a sample of the net, rather than a complete census, but it's something we can afford to do regularly for trend tracking. If we're feeling masochistic some time, we may try a "version" message to try to measure the sampling error. :-) It's always going to be difficult to figure out what's running at sites whose news traffic is gatewayed through from other software -- be it mail or NNTP -- that generates its own message-IDs. The super-jumbo-ID problem will be reduced somewhat when we start using a more compact ID format (it's coming), although we do plan to continue making the IDs distinct enough for this sort of analysis. -- It is not possible to both understand | Henry Spencer at U of Toronto Zoology and appreciate Intel CPUs. -D.Wolfskill| henry@zoo.toronto.edu utzoo!henry
tanner@cdis-1.compu.com (Dr. T. Andrews) (08/09/90)
In article <1990Aug7.180710.3872@zoo.toronto.edu> henry@zoo.toronto.edu (Henry Spencer) writes:
) Actually C News isn't doing too badly; based on analysis of
) message-IDs in our history file, C News is running at 600+ sites,
And even this may be mis-leading. We run C news, but I'll be
dipped if my machines are going to generate C news message-IDs
which read like Russian novels. Just look at the references
header!
I'll send the "seq" program which generates these 7-char IDs
(a 32-bit number expressed in base 36) to anyone who applies.
I'll include with the one-line change to the "anne.jones" script.
--
uflorida!ki4pv!cdis-1!tanner {uunet dsinc}!cdin-1!cdis-1!tannermichael@fts1.uucp (Michael Richardson) (08/15/90)
In article <1990Aug7.180710.3872@zoo.toronto.edu> henry@zoo.toronto.edu (Henry Spencer) writes: >In article <26BCDA30.21C@intercon.com> amanda@mermaid.intercon.com (Amanda Walker) writes: >Actually C News isn't doing too badly; based on analysis of message-IDs >in our history file, C News is running at 600+ sites, about 12% of the >sites generating weird message-IDs that aren't obviously either B or C. Hmm. UUPC and AmigaUUCP variants perhaps? Actually, I think most AmigaUUCP's would get counted as BNews. >It would probably be somewhat faster if we were sysadmin-compatible >with B News, because learning new procedures increases the conversion >effort. This bodes ill for any change that involves incompatibility >of a more serious kind. Yes, a lot of this is sheer inertia, as Well, I found the C news batching system slightly incomprehensible, or maybe I couldn't figure out how to do what I wanted... I stuck to 'nbatcher' which I understand quite well. So I'm sort of C news with B+ News (given than nbatcher enhanced B news' sendbatches ) Do the C News batchers strip the 'Xref' or is the receiving system supposed to strip it (if it were a C News system, it would obviously do so). Not really a problem, my downfeed's just found it a little strange.. -- :!mcr!: | < political commentary currently undergoing Senate > Michael Richardson | < committee review. Returning next house session. > Play: mcr@julie.UUCP Work: michael@fts1.UUCP Fido: 1:163/109.10 1:163/138 Amiga----^ - Pay attention only to _MY_ opinions. - ^--Amiga--^
henry@zoo.toronto.edu (Henry Spencer) (08/16/90)
In article <1990Aug14.181725.9003@fts1.uucp> michael@fts1.uucp (Michael Richardson) writes: > Do the C News batchers strip the 'Xref' or is the receiving system >supposed to strip it (if it were a C News system, it would obviously >do so). Not really a problem, my downfeed's just found it a little >strange.. Theoretically, Xref is not supposed to be transmitted. In fact, every news system I am aware of sends it, expecting the destination system to strip, ignore, or rebuild the header. This is mentioned in our notebook/rfcerrata document. -- It is not possible to both understand | Henry Spencer at U of Toronto Zoology and appreciate Intel CPUs. -D.Wolfskill| henry@zoo.toronto.edu utzoo!henry