chuqui@nsc.UUCP (Chuqui Q. Koala) (12/25/84)
> Perhaps we should somehow design this to be at least partially >compatable with the "distributed news" effort? Seems there are some >common traits, albiet on different scales. I don't really see an overlap with the distributed news effort at this point. We are right now trying to decide on a set of primitives needed to support a reading interface-- eventually I hope to see us implement these primitives in a real environtment to see how they work. What we are not doing right now is anything involving the database or the transmission of articles between systems-- all of the functions of uucp, inews, rnews or the batching subsystems. We are simply trying to find some way of building an interface so that rn, vnews, readnews, or whatever can use a set of function calls to do their work without worrying about what is on the other side of the functions-- it could be a library interface for a single machine, a client/server on TCP, a distributed file system or something we can't even imagine now. What we are trying to do is isolate the reading programs from making assumptions about the database so that we can install an interface between the database and the reading programs for whatever environment is appropriate without changing either end. chuq -- From the ministry of silly talks: Chuq Von Rospach {allegra,cbosgd,decwrl,hplabs,ihnp4,seismo}!nsc!chuqui nsc!chuqui@decwrl.ARPA Deadbone erotica is the prickly panic of forgotten milleniums, it is the moldy billion year madness that creeps deep along the spinal behind of my mind.
eric@milo.UUCP (Eric Bergan) (12/27/84)
What I meant to convey (and apparently failed) was that a large part of the distributed netnews talk has centered on a protocol between a client and a server. It seems to me that this may ba analagous to the protocol needed between a remote site and the "new repository" for requesting missing articles. I realize that there are substantial differences, I am merely wondering out-loud (?) if the common parts of both protocols (SEND-ME article 133 in net.wizbang) should be the same. -- eric ...!seismo!umcp-cs!aplvax!milo!eric
mark@cbosgd.UUCP (Mark Horton) (12/28/84)
If you're thinking of using something like this for requesting old articles, be sure to address security. Right now, netnews is send-oriented. Each site, when it gets an article, decides who it wants to send it to and sends it to those sites. The receiving site can throw away some things it gets, but it can't request extra information. If you allow sites to request news by newsgroup or message-id, the possibility exists for these sites to somehow discover the newsgroup name or message-id of private, locally distributed articles that they aren't supposed to get, and ask for them. Mark