FM> Hi, Keith.

KR>  DM>  On the 23/06/96, Bob waffled to Bill..

KR> i thought you might be interested in this article from
"reseller news". the
KR> prices quoted seem to indicate that it may be a month or so old.

 FM> [...]

KR> ps. i didn't type all that in, it is a virtually uneditted output from
KR> omniscan pro (i removed a few excess c/rs). ocr technology has come a long
KR> way. the original was in 3 columns, the software has even re-assembled
KR> hyphenated words!

 FM> Thanks for adding that comment. I was thinking as I was reading it
 FM> that
 FM> it must have been scanned, from the nature of the errors ('e' for 'c'
 FM> for example). There were a few extra 0s in there, what caused those?

 FM> I'm extremely impressed with that OCR example.

so am i, attempts in the past have been pretty dismal. if i had spent
another 30-40 minutes, it would have been near perfect, there is a
correction mode where the software will highlight any words not in it's
dictionary for correction. it suggests words, and shows an enlargement of
the original scan at the same time. the only thing that i did to improve on
a fully automatic scan was to outline the areas to be scanned myself, if
seems to work much better than letting the software do it, and it only
takes a minute or so. you can also improve the output quality by typing in
some text then scanning the original, it will learn the peculiarities of
the font in use. 

i think that the extra "o"s are crap picked up from between the
columns. the original was basically high quality newsprint.

                        keith
@EOT:

--- WinPoint 0.2.07 Alpha