A client wants to know why an existing highly complex and expensive existing system is "malfunctioning".
It is given truncated title of data and matches them again full titles available in a local database and returns them before further processing.
Client: Problems occur when similar sounding titles are shortened and then become the same name. The system should scan the whole title to prevent this, but I suspect the software is not too clever.
Me: Erm, how will the system determine the "difference" between the same titles?