Re: Geographic proximity determination...

From: Steve Hamlin <hamlin_at_globalnet.co.uk>
Date: Sat, 23 Sep 2000 00:37:59 +0100
Message-ID: <8qgr22$8ne$1_at_gxsn.com>


This sounds fun actually - how about
* index of districts xref'd with locations * some form of phonetic index of similar locations to that input (UK examples Horley, Chorley & Chorley-le-woods) cross referencing phononym (is this a real word?) with possible matches (eg phononym for Boston might xref to Boston(Lincolnshire), Boston(US), Buxton(Derbyshire) * index of locales or districts cross referencing all adjacent districts

Idea is,

1) look for location in district else
2)   derive phononym and lookup SimilarLocations
3)      check for the SimilarLocations within given district else
4)         check for given location in adjacent districts else
5)            check for SimilarLocations in adjacent districts else
6)                 for each adjacent district, check its adjacent districts
.......

Any use? Obviously details depend upon users familiarity with area, expected quality of input, granularity required (is it apartment 2B or 28?) and geographical spread

Steve Hamlin
www.searchandmatch.com

"Ray" <iguana_at_one.net> wrote in message news:39C351CE.46D00FA5_at_one.net...
>
> Howdy all...
>
> I've got a design problem...
>
> I need to create a database of address information. When this
> database is searched for a location in, say Loveland, if the location
> doesn't exist in Loveland but a similar one exists in Indian Hill it
> should suggest it before it suggests a location in, say, Fort Thomas, KY
>
> (which is further away from Loveland than is Indian Hills for all you
> non-Cincinnatians...).
>
> Does anyone have any clue how to go about getting this information
> or encoding it in a database in a reasonable fashion?
>
> I was thinking something about begin able to use a coordinate system,
> but that seems like a pain... I'll be using a MySQL database and PERL
> to design the site... So just about any acrobatics to get the
> information is do-able...
>
> Any insight on this problem is greatly appreciated.
>
> Ray.
>
Received on Sat Sep 23 2000 - 01:37:59 CEST

Original text of this message