Library Lawsuit

Tara Calishain calumet at mindspring.com
Thu Jun 4 09:14:02 EDT 1998


>By the way, helping a student search for information on Bonnie and Clyde,
>Dillenger , and Capone, we first pulled up a German web page Bonnie and Clyde
>a picture of a guy with a gun and a picture of a girl with a gun.  Then the
>JAVA Applet started that began a slideshow of graphics.  The second picture
>had Bonnie and Clyde in bed with the covers pulled up to their chins.  By
this
>time I was already grabbing the mouse trying to hit the back button quickly.
>I told the student "There's no telling what you will find on an internet
>search".  It seems that there were several links to this German porn site
>referencing the "series" of Bonnie and Clyde.  Even a clever play on words
>isn't necessary to get porn sites.  Ever check the source code for one of
>these pages.  How are you going to filter a site that sets keywords half a
>page long of unrelated words that the search sites pick up automatically with
>robots?

Incidental pornography while doing Internet research can be a real problem.
However, I have a couple of suggestions that might help. 

1) Go to a natural language site and ask from there. Asking The Electric Monk
"Who were bonnie and clyde?" brought up far more relevant results than it did
porn -- and the relevant results were more obvious from their summaries. If
that
still makes you uncomfortable, Ask Jeeves has a kid's version at 
http://www.ajkids.com/ . If Jeeves himself doesn't know the answer to your
question, he does a metasearch that's filtered through Surfwatch.

2) Use the domain: syntax that Alta Vista and HotBot offer. For example, this
AltaVista search phrase:
+Bonnie +Parker +Clyde +Texas +domain:us
Forces the search results to originate from .us domains, which are places like
schools and government sites. The first result was a student's report on B&C,
and a lot of the other 67 results were from .tx.us, which of course is Texas.

If you were doing more scientific/professional research, you could use the
+domain:edu request and force your results to come only from edu sites. 

Conversely, you could use -domain:com and filter out all the sites from a com 
domain, but that still leaves .net domains and foreign domains. 


I hope these ideas help.

Tara 




Tara Calishain                  calumet at mindspring.com
_Official Netscape Guide to Internet Research_
http://www.coppersky.com


More information about the Web4lib mailing list