OCR

[Leo Mauler]

--- bewkard <bewkard at gmail.com> wrote:

> I have finally had it with paperwork.  This last 
> tax season did me in.
> 
> I've talked to a couple people about using OCR 
> to store documents digitally.  I know that a 
> few people on the list do this as well.  I was
> wondering if anyone could give me some tips 
> about what works and what doesn't work.  Is it 
> better to OCR things?  is it better to scan and 
> save a PDF or some other portable document?

Unless the document is printed in an OCR-friendly
font, you aren't going to have a great deal of success
with even modern OCR software.  If all you need is to
replace a print copy with a visual image, far better
to scan the document as a graphic image and then store
the graphic image in some graphics format.  

If the documents need to be viewed by other people
then PDF is a good choice, but if you are going to be
the primary viewer of the documents all you really
need to do is scan into a graphics format (like EPS,
TIFF, or an application-specific but open format like
the GIMP's XCF format) and save those files as-is.

Hardcopy tends to outlast digital storage methods, so
some companies are using OCR in reverse (printing
documents to read in with OCR much later on) to store
some of their more long-term information.  They are
making things very easy for the computer though,
because they are using print fonts which are very easy
for OCR applications to read.

It costs a lot of money to get a computer to
accurately extract information from a printed surface,
as the scientists who extracted the earliest recording
of the human voice (from a graphite-sheathed cylinder)
discovered themselves:

French folk song is 'world's earliest recording',
beating Edison by 11 years
March 27, 2008

http://tinyurl.com/34ac5t

Their experience is not entirely unlike that of trying
to use modern OCR software.


      ____________________________________________________________________________________
You rock. That's why Blockbuster's offering you one month of Blockbuster Total Access, No Cost.  
http://tc.deals.yahoo.com/tc/blockbuster/text5.com