I think I'd make sure I knew why I wanted to OCR the images before
making the investment of time and effort. If you just need a really
rough approximation of the text and hope that it picks up some
keywords, most desktop OCR systems will do the job well enough -- and I
would create PDFs, simply because those will work better with a
browser-based retrieval system.

If you really want accurate full text, you might be better off
capturing the documents as images, then finding a vendor who will rekey
the documents for you (preferably using a double-keying entry process
for accuracy).

As noted, be aware that if the documents are not in great shape, you're
going to have a fairly slow capture process, since you won't be able to
use ADF. That will add considerable cost and greatly slow the process.
If you plan to use a vendor, make sure they see your worst documents
and are willing to commit to you that they won't tear up your documents
when they are scanned (assuming that you want the originals back).

An alternative is to capture the documents, then work out an indexing
process that captures a series of keywords that are relevant to your
organization and would help you find the needed document. A key field
is the date of the release -- regardless of which method you choose,
that field will need to be accurate.

Patrick Cunningham, CRM

List archives at http://lists.ufl.edu/archives/recmgmt-l.html
Contact [log in to unmask] for assistance