Compile OCR Tessaract on Kubuntu

As some of you may know, I am trying to scan a whole book written by my grandfather, as I plan to publish it on the internet. It is a very inspiring book which could change our society.

So, first thing first, I need to scan the whole book and use an Optical Character Recognition (OCR) software as a first step to put it into HTML form.

This is the main documentation wiki for Ubuntu OCR.

I only had a test scan in .pdf format to test with, so I followed the instructions to use GIMP to transform it into a TIFF (.tif) file.

I used Tessaract to OCR the resulting image but got an empty file as a result.

I found this issue
which was fixed in tesseract 2.04, which is a higher version from the ubuntu one. So I had to compile it from source.

I downloaded the 2.04 source, and the french and english language files.
I tried ./configure but got:

Apparently, I was missing the g++ compiler.
I installed the package build-essential and tried again ./configure which worked this time.

I then tried make which ended with the following error:

It was a known issue which had only been fixed in the upcoming 3.0 version.

The solution:
edit "viewer/svutil.cpp"
And... add in the line below at about line 35:

Note that the ubuntu package installs the tessaract binary here: /usr/bin/tesseract, while the manually compiled software is at: /usr/local/bin/tesseract. Make sure to use the full path to call the version you want.

Unfortunately, compiling tessaract didn't help me. I need to do a better scan of the book, directly to TIFF format this time.
I still posted this here, hoping it would help someone. If it did, do post a comment below. Spammers may abstain: comments are moderated.