Cameron Laird's personal notes on DéjàNews

DejaNews [DN] named the original "best-known and largest database of Usenet news postings". While there has not been a DN since May 1999, I use it here to abbreviate any service useful for searching Usenet archives, and especially those which can yield URLs to specific posts suitable for sharing with colleagues.

Suppose one is looking at a Usenet article. How does one synthesize a useful URL? One way is with formulas such as http://groups.google.com/groups?ic=1&selm=1ee2tc3.vy63uq1gycw8sN@207-229-150-45.d.enteract.com, an example of the form whose crucial element is selm=message-ID. An alternate spelling for the same (?) is http://groups.google.com/groups?ic=1&q=msgid:<messageID>.

Google has apparently set its interface to be most succinct for display of threads, of which http://groups.google.com/groups?ic=1&th=e01376f9a4ca5501 is an example.

Old DN-style article numbers can be used in May 2001 to bring single articles into view. http://groups.google.com/groups?ic=1&selm=an_650092068 is an example of such a formula. Note the "650092068" here is a DN article number.

Jorn is working on a tutorial on URL hacking.

What follows hasn't yet been updated to the Google era. I leave it here for now in anticipation of eventual re-write and consolication.


Introduction to DejaNews

[...]

[To-do: tie this page in with the indexes I maintain on newsgroup archives. Also, explain to newcomers how to reach deeper history with DN; I get a lot of requests for that.]

[Explain alt.fan.dejanews.]

DejaNews Digesters

Several people have prepared front ends to DN's search facilities, generally to scrub advertising and other noise from the results. Several are designed for local hosting. Most are command-line utilities coded in Perl. These include Re:News apparently is a "Newsreader that searches for articles in Usenet newsgroups", that is, a commercial front-end to the Deja data-store.

There are also a few Web sites which wrap DejaNews; that is, they offer superior interfaces for searching, and/or display results more conveniently. The raw work of searching they continue to hand off to DejaNews, for they don't maintain their own archives. The examples I know in this category are

[Explain WWW::Search, search, AutoSearch, WWW::Search::DejaNews.]

[Explain others. Explain plans for future.]

Formulae

DejaNews makes no guarantees about what we can regard as its internal API. However, those of us who've experimented with it have come up with several formulae, or grammars, for URLs which are generally honored. Consider, for example, Usenet posting <378a3f86.28748632@nntp.uio.no>, in which Alf Christopherson relates personal experiences in backyard fruit cultivation in Bergen, Norway. DejaNews assigns an internal ID of "500218176" to this posting, and makes it available as Several variations on the =dnc (DejaNews Classic) element in a URL deserve mention:

[explain unreliability--works for some, not for others; still narrowing down conditions] [explain DN customer non-support]

[Explain "[ST_rn=ps]" formulae.]

Jeffrey Hobbs rightly points out that all the '['-containing URLs are non-conformant; '[', according to RFC XXXX, requires encoding as "...=dnc/%5b%5d/..."

[x$NN vs. www vs. search] [synthesize grammar]

Sources for information about DejaNews

[Explain.]

Hidden goodies

Thanks to Jorn Barger for reporting the hierarchy browse URL.

Kyler Laird whipped up a formula in about 100 seconds that reports on, for example, Guido van Rossum's most recent postings. Kyler cautions that, as a "quick kludge ... it might/should change."

Editorial

[Explain Deja's wicked wastefulness in frustrating us.]

Acknowledgments

Thanks to Jorn Barger, Asher Blum, Geoff Eldridge, Alexandre Ferrieux, Otis Gospodnetic, Jonathon King, Kyler Laird, Larry Virden, and Jean-Claude Wippler, among others, for help researching this page.
Cameron Laird's personal notes on DejaNews/claird@phaseit.net