IRS search/AltaVista

d scott brandt techman at omni.cc.purdue.edu
Wed Apr 23 11:01:12 EDT 1997


Bunch of search heuristic aspects at play with search engines on the 
Internet... Nick is right on target with his observation of things being 
a "trade-off" especially given how the tools index and rank pages. But if 
you do a search like  +"Internal Revenue Service" +url:gov  for instance, 
you'll find a lot of pages with No Title-- one of the key fields used for 
indexing and relevancy ranking. Hell, some have *no* HTML tags at all.
Argh! There oughta be a law: "No shirt, no shoes, no <tags>-- no indexing."

Scott
+++++
D. Scott Brandt				  Technology Training Librarian
Assistant Professor		 	    Purdue University Libraries
<techman at purdue.edu>	       <http://thorplus.lib.purdue.edu/~techman>	

On Tue, 22 Apr 1997 Nick Arnett <narnett at verity.com> wrote:

>AltaVista certainly is state-of-the-art in terms of its size, but it is a
>bit below the state-of-the-art in search accuracy.  I hope no one assumes
>that this is anything other than a trade-off.  There are more accurate
>search tools that would handle such a query better.  Of course, even the
>best search tools still make significant errors in accuracy. 





More information about the Web4lib mailing list