lear@turbo.bio.net (Eliot) (07/23/90)
It surprised me that Supersedes is had not been widely implemented. It was initially hoped that the usage of Supersedes in the map data would encourage people to move to the new software before their spool directories filled. That didn't happen, because most of the uuhosts like scripts copy off the maps from the spool, with expiring deleting them right behind uuhosts. Amanda may have a very good point about creating a ``USENET II''. -- Eliot Lear [lear@turbo.bio.net]
fmayhar@hermes.ladc.bull.com (Frank Mayhar) (07/24/90)
In article <Jul.22.15.15.17.1990.726@turbo.bio.net>, lear@turbo.bio.net (Eliot) writes: |> It surprised me that Supersedes is had not been widely implemented. |> It was initially hoped that the usage of Supersedes in the map data |> would encourage people to move to the new software before their spool |> directories filled. That didn't happen, because most of the uuhosts |> like scripts copy off the maps from the spool, with expiring deleting |> them right behind uuhosts. Well, if you didn't have to unshar the map files, we could just leave the maps in the spool directory, and Supercedes would work like as originally expected. Unfortunately, you _do_ have to unshar them, and it's way too expensive to keep two copies of the maps. So...unshar the maps into some other area as they come in, and quick-expire the map articles. I would personally like very much to be able to just feed the articles themselves into pathalias (and I probably could if I weren't too lazy to write a script to do it), but the way the maps are currently handled doesn't lend itself well to that. |> Amanda may have a very good point about creating a ``USENET II''. I agree. That's probably the only way we'll be able to make significant changes in anything like a timely manner. -- Frank Mayhar fmayhar@hermes.ladc.bull.com (..!{uunet,hacgate}!ladcgw!fmayhar) Bull HN Information Systems Inc. Los Angeles Development Center 5250 W. Century Blvd., LA, CA 90045 Phone: (213) 216-6241
blm@6sceng.UUCP (Brian Matthews) (07/24/90)
In article <1990Jul23.204531.3178@ladc.bull.com> fmayhar@hermes.ladc.bull.com writes: |Well, if you didn't have to unshar the map files, we could just leave the maps |in the spool directory, and Supercedes would work like as originally expected. |Unfortunately, you _do_ have to unshar them, and it's way too expensive |to keep two copies of the maps. Nope. At my previous job, I had modified pathalias to skip the shar header stuff and then process the map info, so I didn't need the map files except in the spool directory. Unfortunately I lost the changes when I changed jobs, but as I recall they were fairly simple. I'd make the changes again, except I now just punt path lookup to a neighboring site :-) (with the sites permission, of course.) ||> Amanda may have a very good point about creating a ``USENET II''. |I agree. That's probably the only way we'll be able to make significant |changes in anything like a timely manner. That's the only way we'll be able to make any significant changes *ever*. Usenet (Classic :-)) is very entrenched and many of the installations unsupported, so any change that requires all sites to upgrade their software will never fly. Usenet II (Usenets? Usenetter? :-)) would, at least initially, be completely separate (but maybe parallel), and any site participating would have to be willing and able to upgrade their software fairly quickly, or be dropped. Eventually gateways could be written, but I don't think compatibility should be a consideration for Usenet II.