coolidge@brutus.cs.uiuc.edu (John Coolidge) (09/18/89)
After some argument, it appears that the map entry for brutus (brutus.cs.uiuc.edu) will not be carried in the Usenet maps, at least for now. The argument given is that, since brutus is currently an nntp-only site, that there is no point in carrying brutus in the maps, since they give pathalias data and we don't need any of that. On the other hand, there is already at least one NNTP-only site in the maps (rpi.edu, u.usa.ny.3). There may be more. There seems to be some value seen by the sites involved in having a listing in the maps, even if no UUCP information needs to be carried. Here are the advantages I see in carrying all Usenet sites in the maps who wish to submit an entry: 1) The list of contacts is maintained. One of my primary uses of the maps is to get contact data when sending to a given site. 2) News propagation can be computed. With proper use of the #U field, it's much easier to discover where the news links are. 3) The paper maps showing traffic flow, sites, and connectivity. Without an entry in the maps there's no place to find location or even existance for some sites. My location is obvious (hopefully), since there are a couple other sites in the same building, but for some locations it may not be obvious at all. 4) Census information. The maps provide an easy way of getting a count of sites on the net, the equipment types involved, etc. 5) Announcements of connection availability. Several sites list 'we are willing to take <xxx> type connections' or 'connections in the <xxx> area', etc. 6) Advertisements :-) 7) Providing junk mail addresses :-) The costs that I can see are: 1) Potentially cluttering the namespace listing additional sites. 2) Added size of the maps. Neither of these two costs appear to me to be much of an issue. Not all that many additional sites will want to be added, and many of them will be domainized. The added size of the maps is hardly an issue: they're carrying useful information, and as long as that function is being carried out, volume arguments shouldn't stop things. I suppose the bottom line is: are the maps Usenet maps, or are they UUCP maps? My contention, based on both usage and the types of information that they carry, is that they're de facto Usenet maps, and they should be de jure Usenet maps as well. Comments? --John -------------------------------------------------------------------------- John L. Coolidge Internet:coolidge@cs.uiuc.edu UUCP:uiucdcs!coolidge Of course I don't speak for the U of I (or anyone else except myself) Copyright 1989 John L. Coolidge. Copying allowed if (and only if) attributed. You may redistribute this article if and only if your recipients may as well.
tneff@bfmny0.UU.NET (Tom Neff) (09/18/89)
The UUCP maps should be just that, information necessary and
sufficient to keep UUCP running. That's why the contacts etc.
are listed.
Desirable as it may be to keep a database of all the other
machines in the world, so that we can look stuff up when we
want to, I think that should be handled independently,
albeit using the UUCP map data as one component. LISTSERV
could do this -- just send it a site name and get back a
descriptive entry.
--
'We have luck only with women -- \\\ Tom Neff
not spacecraft!' *-((O tneff@bfmny0.UU.NET
-- R. Kremnev, builder of FOBOS \\\ uunet!bfmny0!tneff (UUCP)