Need advice on spell-checking for city names

[Marc Bissonnette]

Hi all;

I'm hoping someone can point me in the direction of an FAQ or other
appropriate reading material in order to learn how to solve my problem.

I have users submitting multiple city names in a perl application and this
often results in multiple spellings, depending either on their local
dialect, mis-spellings, or copy-and-pasting from a micro$oft app.

For example, the city of Montréal has appeared as

Montréal
Montreal
MontrÈal

Or I'll get close-but-not-correct spellings, like

Ottawa (correct)
Otawa
Ottawaa
Autawa

etc.

How can I go about reducing the numbers of these incidences ? Providing
menus of all possible cities is not feasible, since there are multiple
areas within each province that each need a city list...

Many thanks in advance for insights.

 

[Marc Bissonnette]

Marc Bissonnette ([email protected]) wrote on MMMDCCLXXIII
September MCMXCIII in
<URL:news:[email protected]>: [] Hi
all; []
[] I'm hoping someone can point me in the direction of an FAQ or
other [] appropriate reading material in order to learn how to solve
my problem. []
[] I have users submitting multiple city names in a perl application
and this [] often results in multiple spellings, depending either on
their local [] dialect, mis-spellings, or copy-and-pasting from a
micro$oft app. []
[] For example, the city of Montréal has appeared as
[]
[] Montréal
[] Montreal
[] MontrÈal
[]
[] Or I'll get close-but-not-correct spellings, like
[]
[] Ottawa (correct)
[] Otawa
[] Ottawaa
[] Autawa
[]
[] etc.
[]
[] How can I go about reducing the numbers of these incidences ?
Providing [] menus of all possible cities is not feasible, since
there are multiple [] areas within each province that each need a
city list...


You might want to take a look at the Text::Soundex or the
String::Approx module.

Many thanks, Abigail - will start reading both of those right away.

Or you could ask for a zip code instead of a
city name.

That would be the logical thing to do, yeah, but with this particular
application, it needs to cater to the LCD in what people find most
comfortable (and given that the services it looks up are not always
contiguous throughout some of our rather vast geographical postal code
areas, I had to nix it