Fellow members,

I'm looking for some ideas/suggestions on how to handle a task.

I'm trying to develop a process that will delete from our server the user
id of transferred/terminated employes.  On a regular basis I'll be getting
from HR a list of employes and using a derivation of their payroll number
I'll be deleting the corresponding user id.

The problem that I'll be facing stems from the fact that our users come
from a variety of organizations (each organization could have originated
from a different company that was acquired by Raytheon), and,
unfortunately, there may be payroll number duplication (While a new unique
payroll number is available, due to historical reasons and field size, most
of our user ids are based on the 'legacy' numbers).

The problem can be avoided if a further check of the employe's name is
performed.  However there is no guarantee that the spelling of the name
will be consistent, in the input file and our security database (Users
could have been added by different application managers).

So how does one perform an good approximate match on the user's full name
when one is not sure on the order of 'first' and 'last'  name, nor the
correct spelling of each?

I've been looking into using a 'soundex' coding scheme, and was wondering
if anyone had any better ideas.

How do spell checkers provide a 'list' of possible choices?

Regards
Paul Christidis

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *