Merge pull request #35 from jan-ruzicka/patch-1

more link updates

README.md

@@ -152,5 +152,6 @@ Having said that, in 1995, this engine was in the top 3 in terms of character
 accuracy, and it compiles and runs on both Linux and Windows.
 As of 3.01, Tesseract is fully unicode (UTF-8) enabled, and can recognize 39
 languages "out of the box." Code and documentation is provided for the brave
-to train in other languages. See code.google.com/p/tesseract-ocr for more
-information on training.
+to train in other languages. 
+See [Tesseract Training wiki](https://github.com/tesseract-ocr/tesseract/wiki/TrainingTesseract) 
+for more information on training. Additional [code and extracted documentation](http://tesseract-ocr.github.io/) was generated by Doxygen.