Databases: was Meeting-House

Dave Wilton dave at WILTON.NET
Fri Nov 15 18:54:23 UTC 2002


> Approx. 16 of the installments seem to be missing.
> Makes me wonder what else isn't there.

I've noticed with the MOA database, much of the problem is with the scanning
and optical character recognition of the texts. The information is there and
is correct in the graphical representations of the pages, but the search
engine works off the scanned text, and the optical character recognition
makes a lot of errors. You can see this easily with MOA because that service
allows you to look at both the scanned text and a picture of the original.

For example, I was searching for "Y'all" and kept getting hits for "Wall"
(very frustrating). The OCR had transliterated W into Y' in many instances.
(Hint: If you are searching MOA for a word that begins with W, try searching
for it as Y')

I would bet that many of the other databases suffer from a similar problem.



More information about the Ads-l mailing list