ocr not working 100% yet?

Hi all, yesterday i noticed a document i scanned (scansnap 510), and i had ocr’ed in devonthink wasn’t converted to text completely. Actually the 2 most important parts of the letter where missing; the name of the sender, and the document number…! I took the original scan and threw that into acrobat cs4 and that put a 100% ocr over it; everything turned out fine.
I use the latest beta, anyone know what’s going on?

And is there an easy way to re ocr an entire database?

Thanks a lot, gerben.

For the record, beta 4 solved this!