http:^^www.cs.washington.edu^research^projects^softbots^papers^metacrawler^www4^html^overview.html

来自「This data set contains WWW-pages collect」· HTML 代码 · 共 1,261 行 · 第 1/4 页

is to use provider-created interfaces. Providers could create
an interface for the MetaCrawler to access their service which, in
addition to returning relevant hits, also 
returns the appropriate advertisement. Another solution involves the
MetaCrawler accepting advertisements, and forming a
profit-sharing relationship with the service providers. We are
currently investigating these and other methods of mutually beneficial
co-existence with service providers.
<P>
<A NAME=conclusions></A>
<H2><A NAME=SECTION00060000000000000000> Conclusions</A></H2>
<P>
In this paper we have presented the MetaCrawler, a meta-service for Web
searching with additional features designed to return more references
of higher quality than standard search services. We demonstrated that
users follow references reported by a variety of different search
services, confirming that a single service is not sufficient (Table
<!WA58><!WA58><!WA58><!WA58><A HREF="#followedtable">2</A>).  Further, due to the expressive power of the
MetaCrawler's interface, the MetaCrawler was able to automatically
determine that up to 75% of the hits returned can be discarded. Finally, the
performance benchmarks and usage logs also show that the features
provided by the MetaCrawler are both reasonably fast and actually used
in practice.
<P>
The MetaCrawler provides a ``Consumer Reports'' of sorts for Web searchers.
The individual service data extracted from the MetaCrawler's
logs is compelling evidence concerning the quality of each service. By
comparing services using the same query text and recording what
links users follow, we are able
to evaluate the services from a user's point of view. As far as we
know, we are the first to quantitatively compare the
search services used by MetaCrawler on a large sample of authentic user
queries.
<P>
While it is possible that some MetaCrawler features could be
integrated into the search services, others are intrinsic to
meta-services.  By definition, only a meta-service can provide the
coverage gained by using multiple services. Also, as argued earlier,
client-side meta-services can offer user and site customizations, and
absorb the load caused by post-processing of search results. Finally,
there are some features that do not belong under control of search
services for purely pragmatic reasons. For example, as more commercial
search services become available, tools will emerge that select which
services to use on the basis of cost.  An impartial meta-service such
as the MetaCrawler avoids the conflict of interest that would arise if
such a tool were offered by one of the commercial services.
<P>
New Web services are constantly being created. As the number and
variety of services grows, it is natural to group existing services
under one umbrella.  The MetaCrawler goes further than merely
organizing services by creating an integrated <em> meta-service</em> that
moves the interface (and the associated computational load) closer to
the user.  We believe that this trend of moving up the information
``food chain'' will continue.  The MetaCrawler is one of the first
popular meta-services, but many more will follow.
<P>
<H2><A NAME=SECTION00070000000000000000> Acknowledgments</A></H2>
<P>
The research presented in this paper could not have been accomplished
without the help of many individuals.  We would like to thank Mary
Kaye Rodgers, for editing assistance and for putting up with late
nights. Ruth Etzioni and Ellen Spertus provided comments on an earlier
draft.  Dan Weld, Rich Segal, Keith Golden, George Forman, and Donald
Chinn were very vocal and active in testing the early prototypes of
the MetaCrawler, and Craig Horman and Nancy Johnson Burr were
extremely helpful and patient in dealing with it when it ran
amok. Lara Lewis was very helpful in finding references upon
demand. The Internet Softbot group provided early insight into
desirable features of the MetaCrawler, and Brian Bershad and Hank Levy
contributed ideas relating to the impact the MetaCrawler could have on
the Web. Ken Waln aided in early development for his form patches to
the WWW C library, and Lou Montulli helped in later development by
unlocking the secrets of <tt> nph</tt>-scripts and Netscape caching.
MetaCrawler development was supported by gifts from US West and
Rockwell International Palo Alto Research. Etzioni's Softbot research
is supported by Office of Naval Research grant 92-J-1946 and by
National Science Foundation grant IRI-9357772.

<P>
<P><A NAME=SECTIONREF><H2>References</H2></A><P>
<DL COMPACT>
<DT><A NAME=wwwinfoseek><STRONG>1</STRONG></A><DD>
InfoSeek Corporation.
 InfoSeek Home Page.
 <BR> URL: <!WA59><!WA59><!WA59><!WA59><A NAME=tex2html17 HREF="http://www.infoseek.com"><tt>
  http://www.infoseek.com</tt></A>.
<P>
<DT><A NAME=wwwallinone><STRONG>2</STRONG></A><DD>
William Cross.
 All-In-One Internet Search Page.
 <BR> URL: <!WA60><!WA60><!WA60><!WA60><A NAME=tex2html18 HREF="http://www.albany.net/~wcross/all1srch.html"><tt> http://www.albany.net/&#126;wcross/all1srch.html</tt></A>.
<P>
<DT><A NAME=wwwsavvy><STRONG>3</STRONG></A><DD>
Daniel Dreilinger.
 Savvy Search Home Page.
 <BR> URL: <!WA61><!WA61><!WA61><!WA61><A NAME=tex2html19 HREF="http://www.cs.colostate.edu/~dreiling/smartform.html"><tt> http://www.cs.colostate.edu/&#126;dreiling/smartform.html</tt></A>.
<P>
<DT><A NAME=dreilingersavvy><STRONG>4</STRONG></A><DD>
Daniel Dreilinger.
 Integrating Heterogeneous WWW Search Engines.
 <BR> URL: <!WA62><!WA62><!WA62><!WA62><A NAME=tex2html20 HREF="ftp://132.239.54.5/savvy/report.ps.gz"><tt>
  ftp://132.239.54.5/savvy/report.ps.gz</tt></A>, May 1995.
<P>
<DT><A NAME=wwwgalaxy><STRONG>5</STRONG></A><DD>
EINet.
 Galaxy Home Page.
 <BR> URL: <!WA63><!WA63><!WA63><!WA63><A NAME=tex2html21 HREF="http://galaxy.einet.net/galaxy.html"><tt>
  http://galaxy.einet.net/galaxy.html</tt></A>.
<P>
<DT><A NAME=harvesttr><STRONG>6</STRONG></A><DD>
C. Mic Bowman et al.
 Harvest: A Scalable, Customizable Discovery and Access System.
 Technical Report CU-CS-732-94, Department of Computer Science,
  University of Colorado, Boulder, Colorado, March 1995.
 <BR> URL: <!WA64><!WA64><!WA64><!WA64><A NAME=tex2html22 HREF="http://harvest.cs.colorado.edu/harvest/papers.html"><tt>
  http://harvest.cs.colorado.edu/harvest/papers.html</tt></A>.
<P>
<DT><A NAME=wwwharvesthomepage><STRONG>7</STRONG></A><DD>
Michael Schwartz et al.
 WWW Home Pages Harvest Broker.
 <BR> URL: <!WA65><!WA65><!WA65><!WA65><A NAME=tex2html23 HREF="http://town.hall.org/Harvest/brokers/www-home-pages/"><tt>
  http://town.hall.org/Harvest/brokers/www-home-pages/</tt></A>.
<P>
<DT><A NAME=etzioniuicacm><STRONG>8</STRONG></A><DD>
O. Etzioni and D. Weld.
 A softbot-based interface to the internet.
 <em> CACM</em>, 37(7):72--76, July 1994.
 <BR> URL: <!WA66><!WA66><!WA66><!WA66><A NAME=tex2html24 HREF="http://www.cs.washington.edu/research/softbots"><tt>
  http://www.cs.washington.edu/research/softbots</tt></A>.
<P>
<DT><A NAME=wwwyahoo><STRONG>9</STRONG></A><DD>
David Filo and Jerry Yang.
 Yahoo Home Page.
 <BR> URL: <!WA67><!WA67><!WA67><!WA67><A NAME=tex2html25 HREF="http://www.yahoo.com"><tt>
  http://www.yahoo.com</tt></A>.
<P>
<DT><A NAME=javawhitepaper><STRONG>10</STRONG></A><DD>
James Gosling and Henry McGilton.
 The Java Language Environment: A White Paper.
 <BR> URL: <!WA68><!WA68><!WA68><!WA68><A NAME=tex2html26 HREF="http://java.sun.com/whitePaper/javawhitepaper_1.html"><tt>
  http://java.sun.com/whitePaper/javawhitepaper_1.html</tt></A>.
<P>
<DT><A NAME=wwwinfoMarket><STRONG>11</STRONG></A><DD>
IBM, Inc.
 infoMarket Search Home Page.
 <BR> URL: <!WA69><!WA69><!WA69><!WA69><A NAME=tex2html27 HREF="http://www.infomkt.ibm.com"><tt>
  http://www.infomkt.ibm.com</tt></A>.
<P>
<DT><A NAME=wwwqbic><STRONG>12</STRONG></A><DD>
IBM, Inc.
 Query By Image Content Home Page.
 <BR> URL: <!WA70><!WA70><!WA70><!WA70><A NAME=tex2html28 HREF="http://wwwqbic.almaden.ibm.com/~qbic/qbic.html"><tt> http://wwwqbic.almaden.ibm.com/&#126;qbic/qbic.html</tt></A>.
<P>
<DT><A NAME=kosterrobots><STRONG>13</STRONG></A><DD>
Martijn Koster.
 Robots in the Web: threat or treat?
 <em> ConneXions</em>, 9(4), April 1995.
<P>
<DT><A NAME=wwwlexisnexis><STRONG>14</STRONG></A><DD>
LEXIS-NEXIS.
 LEXIS-NEXIS Communication Center.
 <BR> URL: <!WA71><!WA71><!WA71><!WA71><A NAME=tex2html29 HREF="http://www.lexis-nexis.com"><tt>
  http://www.lexis-nexis.com</tt></A>.
<P>
<DT><A NAME=wwwlycos><STRONG>15</STRONG></A><DD>
Michael Mauldin.
 Lycos Home Page.
 <BR> URL: <!WA72><!WA72><!WA72><!WA72><A NAME=tex2html30 HREF="http://lycos.cs.cmu.edu"><tt>
  http://lycos.cs.cmu.edu</tt></A>.
<P>
<DT><A NAME=lycossignidrv><STRONG>16</STRONG></A><DD>
Michael L. Mauldin and John R. R. Leavitt.
 Web Agent Related Research at the Center for Machine Translation.
 In <em> Proceedings of SIGNIDR V</em>, McLean, Virginia, August 1994.
<P>
<DT><A NAME=wwwhomr><STRONG>17</STRONG></A><DD>
Max Metral.
 Helpful Online Music Recommendation Service.
 <BR> URL: <!WA73><!WA73><!WA73><!WA73><A NAME=tex2html31 HREF="http://rg.media.mit.edu/ringo/ringo.html"><tt>
  http://rg.media.mit.edu/ringo/ringo.html</tt></A>.
<P>
<DT><A NAME=wwwcusi><STRONG>18</STRONG></A><DD>
Nexor.
 CUSI (Configurable Universal Search Interface).
 <BR> URL: <!WA74><!WA74><!WA74><!WA74><A NAME=tex2html32 HREF="http://pubweb.nexor.co.uk/public/cusi/cusi.html"><tt>
  http://pubweb.nexor.co.uk/public/cusi/cusi.html</tt></A>.
<P>
<DT><A NAME=wwww3enginelist><STRONG>19</STRONG></A><DD>
University of Geneva.
 W3 Search Engines.
 <BR> URL: <!WA75><!WA75><!WA75><!WA75><A NAME=tex2html33 HREF="http://cuiwww.unige.ch/meta-index.html"><tt>
  http://cuiwww.unige.ch/meta-index.html</tt></A>.
<P>
<DT><A NAME=wwwopentext><STRONG>20</STRONG></A><DD>
Open Text, Inc.
 Open Text Web Index Home Page.
 <BR> URL: <!WA76><!WA76><!WA76><!WA76><A NAME=tex2html34 HREF="http://www.opentext.com:8080/omw/f-omw.html"><tt>
  http://www.opentext.com:8080/omw/f-omw.html</tt></A>.
<P>
<DT><A NAME=wwwpls><STRONG>21</STRONG></A><DD>
Personal Library Software, Inc.
 Personal Library Software Home Page.
 <BR> URL: <!WA77><!WA77><!WA77><!WA77><A NAME=tex2html35 HREF="http://www.pls.com"><tt>
  http://www.pls.com</tt></A>.
<P>
<DT><A NAME=wwwwebcrawler><STRONG>22</STRONG></A><DD>
Brian Pinkerton.
 WebCrawler Home Page.
 <BR> URL: <!WA78><!WA78><!WA78><!WA78><A NAME=tex2html36 HREF="http://webcrawler.com"><tt>
  http://webcrawler.com</tt></A>.
<P>
<DT><A NAME=webcrawlerwww2><STRONG>23</STRONG></A><DD>
Brian Pinkerton.
 Finding What People Want: Experiences with the WebCrawler.
 In <em> Proceedings of the Second World Wide Web Conference '94:
  Mosaic and the Web</em>, Chicago IL USA, October 1993.
<P>
<DT><A NAME=wwwvirtourist><STRONG>24</STRONG></A><DD>
Brandon Plewe.
 The Virtual Tourist Home Page.
 <BR> URL: <!WA79><!WA79><!WA79><!WA79><A NAME=tex2html37 HREF="http://wings.buffalo.edu/world"><tt>
  http://wings.buffalo.edu/world</tt></A>.
<P>
<DT><A NAME=genmagicguidetocodewarrior><STRONG>25</STRONG></A><DD>
Daniel Sears.
 Guide to CodeWarrior Magic/MPW.
 Development Release 1<BR> URL: <!WA80><!WA80><!WA80><!WA80><A NAME=tex2html38 HREF="http://www.genmagic.com/MagicCapDocs/CodeWarriorMagic/introduction.html"><tt>
  http://www.genmagic.com/MagicCapDocs/CodeWarriorMagic/introduction.html</tt></A>, May
  1995.
<P>
<DT><A NAME=wwwdejanews><STRONG>26</STRONG></A><DD>
DejaNews Research Service.
 DejaNews Home Page.
 <BR> URL: <!WA81><!WA81><!WA81><!WA81><A NAME=tex2html39 HREF="http://www.dejanews.com"><tt>
  http://www.dejanews.com</tt></A>.
<P>
<DT><A NAME=wwwsunmultithreaded><STRONG>27</STRONG></A><DD>
Sun Microsystems, Inc.
 Multithreaded Query Page.
 <BR> URL: <!WA82><!WA82><!WA82><!WA82><A NAME=tex2html40 HREF="http://www.sun.com/cgi-bin/show?search/mtquery/index.body"><tt>
  http://www.sun.com/cgi-bin/show?search/mtquery/index.body</tt></A>.
<P>
<DT><A NAME=wwwmoviedatabase><STRONG>28</STRONG></A><DD>
The Internet Movie Database Team.
 The Internet Movie Database.
 <BR> URL: <!WA83><!WA83><!WA83><!WA83><A NAME=tex2html41 HREF="http://www.msstate.edu"><tt>
  http://www.msstate.edu</tt></A>.
<P>
<DT><A NAME=wwwmckinley><STRONG>29</STRONG></A><DD>
The McKinley Group, Inc.
 Magellan: McKinley's Internet Directory.
 <BR> URL: <!WA84><!WA84><!WA84><!WA84><A NAME=tex2html42 HREF="http://www.mckinley.com"><tt>
  http://www.mckinley.com</tt></A>.
<P>
<DT><A NAME=wwwverity><STRONG>30</STRONG></A><DD>
Verity, Inc.
 Verity Home Page.
 <BR> URL: <!WA85><!WA85><!WA85><!WA85><A NAME=tex2html43 HREF="http://www.verity.com"><tt>
  http://www.verity.com</tt></A>.
</DL>

<P>
<H2><A NAME=SECTION00080000000000000000> About the Authors</A></H2>
<P>
Erik Selberg, <em>
  <!WA86><!WA86><!WA86><!WA86><A NAME=tex2html13 HREF="mailto:[e-mail]">[e-mail]</A></em>, <!WA87><!WA87><!WA87><!WA87><A NAME=tex2html15 HREF="http://www.cs.washington.edu/homes/selberg">http://www.cs.washington.edu/homes/selberg</A><BR> 
Department of Computer Science and Engineering <BR> 
Box 352350 <BR> 
University of Washington <BR> 
Seattle, WA 98195
<P>
Erik Selberg is pursuing his Ph.D. in computer science at the
University of Washington. His primary research area involves World
Wide Web search, although he also has interests regarding system performance
and security as well as multi-agent coordination and
planning. In April, 1995 he created the MetaCrawler, a parallel Web
search meta-service. He graduated from Carnegie Mellon University in
1993 with a double major in computer science and logic, and received
the first Allen Newell Award for Excellence in Undergraduate Research.
<P>
Oren Etzioni, <em>
  <!WA88><!WA88><!WA88><!WA88><A NAME=tex2html14 HREF="mailto:[e-mail]">[e-mail]</A></em>, <!WA89><!WA89><!WA89><!WA89><A NAME=tex2html16 HREF="http://www.cs.washington.edu/homes/etzioni">http://www.cs.washington.edu/homes/etzioni</A><BR> 
Department of Computer Science and Engineering <BR> 
Box 352350 <BR> 
University of Washington <BR> 
Seattle, WA 98195
<P>
Oren Etzioni received his bachelor's degree in computer science from
Harvard University in June 1986, and his Ph.D. from Carnegie Mellon
University in January 1991.  He joined the University of Washington as
assistant professor of computer science and engineering in February
1991.  In the fall of 1991, he launched the Internet Softbots project.
In 1993, Etzioni received an NSF Young Investigator Award.  In 1995,
Etzioni was chosen as one of 5 finalists in the Discover Awards for
Technological Innovation in Computer Software for his work on 
Internet Softbots.
<P>
His research interests include: software agents, machine learning, and
human-computer interaction.
<P>

<H2><A NAME=SECTION000100000000000000000>   About this document ... </A></H2>
<P>
 <STRONG>Multi-Service Search and Comparison Using the 
        MetaCrawler</STRONG><P>
This document was generated using the <!WA90><!WA90><!WA90><!WA90><A HREF="http://cbl.leeds.ac.uk/nikos/tex2html/doc/latex2html/latex2html.html"><STRONG>LaTeX</STRONG>2<tt>HTML</tt></A> translator Version 95.1 (Fri Jan 20 1995) Copyright &#169; 1993, 1994,  <!WA91><!WA91><!WA91><!WA91><A HREF="http://cbl.leeds.ac.uk/nikos/personal.html">Nikos Drakos</A>, Computer Based Learning Unit, University of Leeds. <P> The command line arguments were: <BR>
<STRONG>latex2html</STRONG> <tt>-split 0 www4-final.tex</tt>. <P>The translation was initiated by Erik Selberg on Mon Oct  9 17:24:12 PDT 1995<BR> <HR>
<P><ADDRESS>
<I>Erik Selberg <BR>
Mon Oct  9 17:24:12 PDT 1995</I>
</ADDRESS>
</BODY>