Hello
That's most interesting about Inktomi... I have been following what metadata
search engines index and include this on my Comparison of Search Engine User
Interface Capabilities at:
http://lisweb.curtin.edu.au/staff/gwpersonal/compare.html
Hotbot uses author, description and keyword (and even though it uses
Inktomi's search, it doesn't state that it supports DC and actually provides
a guide on how to input data using the metatag style listed below)
AltaVista uses description and keyword
Excite only supports description
Northern Light indexes metadata but doesn't treat it any differently.
They require the following syntax:
<META NAME="Description" CONTENT="...">
etc.
Because of this, work that is underway at my institution is including these
extra tags along with DC. You can check out our template (still in beta)
at:
http://metadata.curtin.edu.au/template.html
Cheers
Gill
__________________________________________________
Gillian Westera
Acting Information & Education Services Librarian
Curtin University Library & Information Service
GPO Box U1987
Perth, Western Australia 6845
Ph. +61 8 9266 7249 Fax: +61 8 9266 3947
http://lisweb.curtin.edu.au/staff/gwestera.html
__________________________________________________
-----Original Message-----
From: Eileen Quam [mailto:[log in to unmask]]
Sent: Friday, 21 September 2001 0:52
To: [log in to unmask]
Subject: Re: Meta-tags and search engines
Inktomi Search Solution (formerly Ultraseek search engine) offers tuning for
Dublin Core tags to increase their relevance ranking. In addition, it can
use DC.subject for its rule-based Content Classification Engine
(hierarchical topic categories). You can set up advanced search queries to
work with DC tags. Check out our Bridges site for more information and
studies on this (http://bridges.state.mn.us) and the State of Minnesota's
search site for an implementation of the Content Classification Engine
(http://search.state.mn.us). I would be happy to correspond privately about
your other questions.
Eileen
Eileen Quam
Information Architect
Minnesota Dept. of Natural Resources
[log in to unmask]
651.297.2341
651.297.4946 FAX
>>> David Davies <[log in to unmask]> 09/20/01 11:36AM >>>
Dear all
I've only recently joined this list and I still have a lot to learn about
metadata in general and Dublin Core specifically but I'd like to ask an
early question if I may.
The content management system we use to store our curriculum web pages
creates DC tags in the HTML page header thus
<meta name="DC.Contributor" content="David Davies">
<meta name="DC.Format" content="text/html">
<meta name="DC.Language" content="en">
etc
Do internet search engines such a Google use these tags in a meaningful way
or is the text contained in the content attribute just lumped in with the
text of the page itself to improve the chance of a 'hit' when a user submits
a search?
More specifically, I would like to know if anyone knows of a search engine
that can explicitly make use of these tags by allowing a complex search
query? Say I have a multilingual site and I want to pull out pages where
DC.Language" = "en". I'd need to be able to construct a search query that
allowed input of one or more DC meta tags as qualifiers. Anyone any ideas if
there's a search engine that does this?
Our content management system allows the user to search based upon the
values of DC meta tags as well as free text in the body of the page but I
would like to use a search engine that indexes a number of sites yet still
retains this functionality. And by number of site I mean hundreds or
thousands, the kind of order of magnitude that makes Google an useful
engine.
Is this a case of off the shelf or do it yourself?
Cheers,
David.
--
Dr David A. Davies
Senior Lecturer
Medical Education Unit
School of Medicine
The University of Birmingham
B15 2TT. UK.
Tel: 44 (0) 121 414 3255
Fax: 44 (0) 121 414 6919
|