[Sigia-l] FW: Kozaru: new search engine

Avi Rappoport avirr at searchtools.com
Fri Aug 20 19:22:12 EDT 2004


At 1:53 PM +1000 8/13/04, Eric Scheid wrote:
>A new search engine company (looking for investors, of course) claims to be
>developing 'a new search technology' - based on the premise that 'people do
>not think in search terms'.
>
>Whilst I agree with their statement that 'search engine technology doesn't
>really work' (most of the time, with qualification), I'm not so sure about
>their lofty aims.

I'm not so sure I agree with this.  I know that keywords are a 
problem, but so is everything else.  The search engines are now doing 
a better job of bubbling the most useful sites to the top (except 
travel and other really competitive fields).  All the other things 
we've tried, including natural-language search, multi-step interviews 
and complicated forms have been worse. Excite's concept search was 
particularly awful (if anyone's still using Excite for Web Servers, 
please *stop* and switch to SWISH-E).  The vast majority of searches 
are simple keywords and it's very difficult to get past that.  And 
according to a new Pew report, "87% of search engine users say they 
find the information they want most of the time." 
<http://www.pewinternet.org/PPF/r/132/report_display.asp>

The problem is that there are many levels and kinds of information 
needs, and some are very badly served by simple keyword queries.  I 
don't want my lawyer or clinician or patent examiner or chemist or 
historian to be satisfied by the results of a keywords search.  But 
if I'm trying to find out what you've posted about search (very 
interesting, btw), it works fine.  It's making the distinction, 
making sure that people understand the limits of the existing engines 
and the usefulness of vertical-search alternatives, that's likely to 
help the most.

That's why I think this is basically an impossible problem to solve 
-- there's just so much nuance and complexity involved in the 
language, much less the underlying questions -- a webwide search 
engine is too ambitious.  It's like the Do What I Mean computer 
interface: I've been wanting that for years.

Avi


-- 
   Avi Rappoport, Search Engine Consultant <mailto:avirr at searchtools.com> 
   Complete Guide to Search Engines for Web Sites and Intranets
             <http://www.searchtools.com>



More information about the Sigia-l mailing list