In response to Diane Kozelka, Sara Davis, and other users of File 349
(WIPO/PCT Fulltext) on Dialog, we would like to assure you that there is
nothing wrong with the file or its data. As has been stated in the
Chronolog articles announcing File 349, MicroPatent produces the data
using an optical scanning process. This process can introduce unexpected
errors into the electronic text. Depending on the quality of the
original document being scanned, smudged characters, ink spots, etc.
will cause the scanned characters to be translated incorrectly. For
example, the word "associated" sometimes appears as "assodated" because
the "ci" characters are not "OCRed" correctly. The problem tends to be
more pervasive in older patents which were scanned with older
technology.
Sara also mentioned problems with identification of the claims section
of records. MicroPatent segments the text into the detailed description
and claims. In rare cases, the field segmentation program may fail to
identify the correct start of the claims section. Incorrect
identification of claims is more likely to be found in older records,
but appears to affect only a small percentage of the database.
File 349 provides tremendous value to Dialog's customers who can now
search the three major patent sources, USPTO, EPO and WIPO, in a single
fulltext search on the Dialog system. A new OneSearch category, PATTEXT
is available for this purpose, as requested by Roy Zimmermann.
Stay tuned for further developments with File 349.
Sophie Hudnut
Dialog, Intellectual Property Content
Email: sophie_hudnut@dialog.com
-- This message (plus any attachments) is confidential and may be subject to lawyer-client privilege. Use without permission is strictly prohibited.
This archive was generated by hypermail 2b29 : Fri Aug 10 2001 - 15:58:21 EDT