bharathi@acsu.buffalo.edu (Chandra Shekara Bharathi) (04/27/91)
Hi!
Our group is trying to automate reading of addresses on mail pieces
for the United States Postal Service. It is envisaged that such 'reading'
and identifying(or generating, as the case may be) the appropriate 9-digit
Zip code [i.e. 5 digits plus the unique 4 digit add-on ] using all the
fields specified on the address, namely the State, city, 5 digit zip, Street
name, apt., etc. would be at the rate of atleast 12 mail pieces per sec.
For this, we have large databases that link up all the above fields.
It thus becomes evident that database searching has to be done at a
very fast rate. We already have a pretty neat thing running. However, we
would like to improve the speed further.
If anyone has come across papers/references on such database issues
involving large (compressed) databases and fast access mechanisms, i'd very
much like to hear from fellow netters.
Please note that everything is on-line and disk accesses are not
feasible.
Please email replies directly to bharathi@cs.buffalo.edu
Thanks in advance.
Chandra
-------------------------------------------------------------------------------
Chandra Shekara Bharathi |
| internet: bharathi@cs.buffalo.edu
Center of Excellence for Document Analysis | bitnet : bharathi@sunybcs.BITNET
and Recognition [CEDAR] |
Dept of Computer Science |
State Univ of New York at Buffalo |
Buffalo, NY 14260 |
(716) 636 2448 uucp: ..!{ames,boulder,decvax,rutgers}!sunybcs!bharathi
--------------------------------------------------------------------------------