Skip to Content.
Sympa Menu

baslinux - [BL] OCR, was Re: Slax review

baslinux AT lists.ibiblio.org

Subject: Baslinux mailing list

List archive

Chronological Thread  
  • From: baslinux AT lists.ibiblio.org
  • To: baslinux AT lists.ibiblio.org
  • Subject: [BL] OCR, was Re: Slax review
  • Date: Mon, 16 Nov 2009 04:17:09 +0000 (UTC)

http://www.eecho.info/Echo/office/linux-ocr
Review of linux ocr programs.
CLI - gocr (jocr), ocrad, tesseract, ocropus
GUI - Clara, Ocre (GTK2), gtk-ocrad

Linux packages has ocrad 0.16 for SW11 (nothing older) which is 148K
package and requires libgcc_s and libstdc++.so.6
gtk-ocrad has different deps - libX1 libXext libdl libm gtk1.2
The firefox dependencies should suffice.


Would not compile for me with glibc 225 and gcc 2.95.3 (SW81)
Use linuxpackages version. I have not tested it.


gocr 0.37 (310K, SW9 or 9.1), 0.42 (SW11) - 300-400K or so
requires libpbm pgm ppm pnm from elflibs

http://keesan.freeshell.org/bl/gocr-048-glibc225.tgz
Latest version, jul09. Dependencies libm libc and ld-linux.
Might even compile with libc5. 350K stripped binary and some documentation and a sample file. For authors and history see the source code package.


gocr 0.48 was very fast but not entirely accurate. lazy la2y Iazy (as in lazy fox). Three errors in 8 lines of very clear text. I don't know what it would do with Czech diacritics. Might require some hand editing. Read the instructions, there are lots. unicode.txt


Tesseract (HP commercial) is from 1995 made open source in 2005, about 3MB
source code. Works on tiff not p?m.

Compiled but running it produced utter garbage for me. I added their
phototest.tif file (converted to pbm) to the gocr package.
Make install did not even install most of the required files so I had to hand copy and it still produced garbage.

------------
Other than OCR, what programs/features are missing in BL2.1 which lead people to use SLAX, peanut, etc? Not counting GTK2 and fancy package managers and desktops.

I have put Puppy on old computers for Windows users because it is easier for them to learn. But it runs slower and needs more memory and speed due to all the 'for idiots' stuff. A lot easier to set up printing (but then cupsd and other things waste memory when not in use).

My big problem is most stuff written since 2000 won't compile with BL2.
Slackware 11 glibc, gcc and X sometimes work better than SW7/8.
Sometimes I can only compile older versions (usually adequate for our older hardware). My newest hardware is older than Slackware 11.

If I experiment wiht the 2.6 kernel, is there some version to stop at?

Sindi Keesan




Archive powered by MHonArc 2.6.24.

Top of Page