Scanning works better than under Windows. OCR sucks rocks.

As far as I can tell there are only two OCR "engines" for Linux, gocr and 
ocrad. Both might work better if I could get them to use a dictionary. I can't 
find any way to get ocrad to use a dictionary at all. Gocr will, but only under 
Kooka (not XSane), and it won't use any dictionary that I have installed, which 
is to say, the OO.o dictionaries. That's because you have to choose the 
dictionary in a drop-down box that you can't type into, and none of the OO.o 
dictionaries are listed.  Plus there is a bug in Kooka that requires you to 
close and restart the program for each page you OCR. Also, there is no "Scan" 
button on OO.o Writer that will launch a scan/ocr utility -- although that is a 
minor kvetch.

I do realize the OCR software is incredibly complicated. Outfits like Aabbyy, 
OmniPage, and others have spent many milliions of dollars on programmer 
salaries to develop their OCR engines. 

One of my big complaints, however, is that, even if the OCR engine worked well, 
you can't get it to work because the GUI is not implemented correctly. In 
Kooka, for example, you can click on a button and sometimes absolutely nothing 
happens. What passes for documentation sometimes doesn't even explain what a 
button is supposed to do. As far as I can tell the programmer just put some of 
the buttons on the interface and never got around to writing the code to make 
them do anything.

So are there any other Linux OCR choices besides ocrad and gocr? (I include 
command line tools, as long as I can figure out how to use them; i.e., they 
come with understandable documentation.) Or are there any commercial OCR 
utilities that will run under Linux? Or should I hire a typist to keyboard 
stuff into the computer for me?

Right now I have used XSane (a very nice utility) to scan the 35 pages I need 
to TIFF files. XSane does a great job. I used TIFF because I know Windows OCR 
utilities can OCR from a TIFF without needing a scanner. So, push come to 
shove, I can move the TIFFs to my Windows computer and finish up there. But 
there must be something that works under Linux and will create text that I can 
use in Writer. Anyone know of anything?

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to