Comparisons of VERONICA Servers (fwd)
Publib Poster
publll at nysernet.ORG
Sun Feb 19 20:04:33 EST 1995
Sender: jpolly at nysernet.ORG (Jean Armour Polly)
Subject: Comparisons of VERONICA Servers
In the current Library Journal, the sidebar to my 2-part article on
VERONICA has been edited so that it bears little resemblance to my original
text.
Most important to the Internet Community is the information contained in
this part of the article, which compares searches done on various VERONICAs
around the world, on the same day.
People who want the entire article may email me, we are soon to put these
articles up for gopher/www/ftp access. Please redistribute as you see fit.
Jean Armour Polly
One fine day in December, 1994, I spent some time with VERONICA servers in
many countries. This should not be referenced as an exhaustive or
authoritative investigation: it is only my finding on a particular day. On
this day, I found that some servers use old indexes, and some censor their
stoplists to prevent searches on certain words.
I began at my home gopher at NYSERNet.org. Our VERONICA server uses the
most current index, so I compared hits at our server with those from other
servers.
I searched on the terms "erotica" and "children" to get the list of hits to
a manageable size.
NYSERNet's VERONICA found only one hit:
Veronica at NYSERNet: erotica and children
--> 1. Newsgroups Nixed: Erotica + Children = Banned (25nov93)
This referred to a 1992 directive at Cornell which temporarily removed the
newsgroups alt.binaries.erotica "whose transmission may constitute
violations of the United States Code regarding sexual exploitation of
minors" until the allegations could be investigated.
Here are my experiences with this search at other public VERONICA servers.
Bergen, Norway-- found the same hit.
Manitoba, Canada-- found the same hit.
MINITEX, University of Minnesota, US-- found the same hit.
SCS Nevada, US-- found the same hit.
Pisa, Italy-- found the same hit.
PSINet, US-- found the same hit.
Tachyon Communications, US-- on 12/11 I found the same hit, although on
12/8 it would not let me search on "erotica" or "binaries".
Here is where it gets interesting.
The gopher at liberty.uc.wlu.edu (Washington and Lee University, Lexington,
Virginia, US) gives access to a huge list of VERONICA servers, although
many of them do not seem very stable.
This site queries the servers at regular intervals to see who is really up,
it says, but even so, many did not respond when I tried them. The servers
are sorted by recent connection response time. Note it also lists which
servers did not respond to the connection request, either because they were
down, or too busy.
Here is what the menu looked like when I connected:
Search gopherspace by document type, using Veronica (WLU)
Search Gopher Menus by VERONICA
--> 1. NYSERNet 1/
2. Tachyon Communications, Florida 1/
3. University of Cologne 1/
4. PSINet, Virginia 2/
5. University of Minnesota 2/
6. University of Pisa, Italy 2/
7. Keio University .......... (Japanese Gopherspace) 3/
8. FONOROLA ................. (Canadian Gopherspace) 4/
9. SUNET, Sweden 4/
10. University of Bergen, Norway 4/
11. America Online 5/
12. University of Manitoba 5/
-------------------------------
14. veronica queries--How to, and FAQ (from Nevada)/
The Servers are Sorted by Recent Connection Response Time
-------------------------------
Other Servers Did Not Connect (as at: Mon 01:26)
18. AARNET ................... (Australian Gopherspace) -/
19. Manchester University -/
20. SURFnet, Netherlands -/
21. University of Nevada, Reno -/
22. University of Stuttgart .. (German Gopherspace) -/
23. University of Texas, Dallas -/
SUNET
Running the search at the VERONICA server at Sweden's SUNET produced some
unusual results:
--> 1. alt.binaries.pictures.erotica.children/
2. alt.binaries.pictures.erotica.children/
The Cornell "newsgroups nixed" post was not there. Both of these hits went
to dead ends where data no longer existed. Possible old index.
America Online
The same thing happened when searching the America Online VERONICA via the
Washington and Lee Univerity (Lexington, VA) gateway. Only the two dead end
hits were found, ostensibly from the old index. "Bestiality" produced hits.
So did "nixed" but not the Cornell one we're looking for. Possible old
index.
Keio University .......... (Japanese Gopherspace)
Returns hits for "children" and only one screen of hits on "erotica". Also
the word "nixed" returns no hits, so the "Newsgroups Nixed" post is not
there. "Bestiality" returns no hits. Possible old index, possible censored
stoplist.
FONOROLA ................. (Canadian Gopherspace)
One, yes, one hit on "erotica." No hits on "erotica" and "children." Many
hits on "Children." "Nixed" returns nothing. Many hits on "sex." No hits on
"bestiality." Possible old index.
University of Koln (Cologne, Germany)
My search on "erotica" produced nothing. But at least it told me why. Here
is the reason: it was a "dubious" search! (Who KNEW??!!) See the quoted
message below:
Explanation of the message "The Word "erotica" is too common and is not indexed"
Your query for "erotica"
contains the word "erotica".
It is excluded from indexing because it is too common in gopher titles
or is not presented because it is used very often in dubious queries.
This veronica service is located at the University of Koeln (Cologne)
in Germany. It uses a special software version of veronica.
Veronica in general:
- Veronica is a service to retrieve gopher titles by search words.
- The data is collected from nearly all reachable Gopher servers in the
world, stored at an FTP server in Nevada and installed at about 10
Veronica servers in the world.
- Many Gopher servers show links to these Veronicas and most
Gopher users can use Veronica.
- Veronica is develod by Fred Barrie and Steve Foster. The local version
is by me. The name "Veronica" means "very easy rodent-oriented net-wide
index to computerized archives".
Differences of this Veronica server to others:
- Logical expressions are evaluated from left to right (others use right to
left)
- No parentheses are allowed (had been used only in 0.5% of the queries)
- More messages are given (errors in queries, time limit, ...)
- Messages are explained
- Maximal 1000 hits are returned
- Not all hits of very frequently used words (like gif) are collected
Data for this veronica service:
- Data collection at UNR (Nevada): about August 3
- Data accessible at FTP at UNR: August 5
- Data installed here: August 11
- Amount of Data: about 1.18 GB, compressed about 290 MB
Statistics of this service:
- About 8 queries for search words per minute (in the mean)
- Several requests for informations per minute, esp. tests of gopher sides,
if this server is up, and explanations of messages
(Heinz Stoewe)
I respect Heinz's VERONICA because he tells me up front that he's censored
it, and he tells me the date of the index he uses. He also explains how he
has tweaked his local VERONICA to be slightly different than the others. I
think this is responsible VERONICA-keeping. Thanks, Heinz. I know your
service uses an old index, I know you've decided what searches are
worthwhile and which ones aren't. Now I can choose to stay away from yours.
Other VERONICA operators don't give me that choice. Their indices may be
just as old and just as censored-- but we just don't know about it. Users--
if you want better Internet resources, speak up. Help find Steve Foster
(Father of VERONICA at Univ of Nevada- Reno) some funding to enable the
"Reference" servers he envisions in the accompanying article.
(JP note: Foster is troubled by the user's inability to tell which servers
use old indices and stoplisted words. He would like funding for a stable
server that could be used as a "reference" server, that would be inviolate
and would allow many connections.)
Jean Armour Polly
co-moderator PUBLIB and PUBLIB-NET, listservs for
public librarians
author "Surfing the Internet"
More information about the Publib
mailing list