Re: Automating Search Interfaces

Michael Kay (M.H.Kay@eng.icl.co.uk)
Mon, 23 Feb 1998 11:46:48 -0000

This is a multi-part message in MIME format.

------=_NextPart_000_0051_01BD4050.B9604F60
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

>I would like to talk about the location of the person making the search =
versus >the location of the product or service provider
=20
Geographic/Spatial queries are a well-researched topic in the =
database literature. Free text retrieval is definitely a weak approach, =
though people attempt it by using thesaurus facilities to represent the =
structure of a gazetteer. In most of the practical systems I have seen, =
spatial query is done using postal codes: the system needs knowledge of =
which postal districts are near each other. (We also use such techniques =
for scheduling the itinerary of service engineers).
=20
>A hotel room is a 'chambre' in french. If I search for a hotel room =
in Italy, I>don't know the word for room in italian...
=20
Multilingual search is well researched and seems to work reasonably =
well. The more difficult problem is to distinguish agencies that can =
book you a hotel room from newsletter articles by people enthusing what =
a wonderful hotel room they were staying in: I think this is why there =
will always be added value in manual categorization and indexing =
services.
=20
Mike Kay
=20

------=_NextPart_000_0051_01BD4050.B9604F60
Content-Type: text/html;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD W3 HTML//EN">

>I would like to talk about the location of = the person=20 making the search versus >the location of the product or service=20 provider
 
Geographic/Spatial queries are a = well-researched topic in the database literature. Free text = retrieval is=20 definitely a weak approach, though people attempt it by using = thesaurus=20 facilities to represent the structure of a gazetteer. In most of the = practical systems I have seen, spatial query is done using postal = codes: the=20 system needs knowledge of which postal districts are near each other. (We also use such techniques for = scheduling the=20 itinerary of service engineers).
 
>A hotel room is a 'chambre' in french. If I = search for=20 a hotel room in Italy, I>don't know the word for room in=20 italian...
 
Multilingual search is well = researched and=20 seems to work reasonably well. The more difficult problem is to = distinguish=20 agencies that can book you a hotel room from newsletter articles by = people=20 enthusing what a wonderful hotel room they were staying in: I think = this is=20 why there will always be added value in manual categorization and = indexing=20 services.
 
Mike Kay
 
------=_NextPart_000_0051_01BD4050.B9604F60-- xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@ic.ac.uk Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ To (un)subscribe, mailto:majordomo@ic.ac.uk the following message; (un)subscribe xml-dev To subscribe to the digests, mailto:majordomo@ic.ac.uk the following message; subscribe xml-dev-digest List coordinator, Henry Rzepa (mailto:rzepa@ic.ac.uk)