>Yes, indeed, the human eye does an awfully good job of that. Your solution is what we did as well.
>
>I was commenting (indirectly) on the statement of "a database of probable first and last names". I'm not sure how such a database would be of value.
>
In our algoritm we did use First Names database (it's not possible to have last name database). We also had a table for suffixes and words that can indicate the organization...
It's quite an interesting application. As I said, in my former job I did couple of them with a somehow similar interface ideas: first the process attempts to do automatic cleanup, then the user cleans up ~2-15% of the bad cases...
If it's not broken, fix it until it is.
My Blog