OCR, or optical-character-recognition, is taking on more and more importance for translation buyers and producers. The reason is that translators themselves usually charge by the word, not by the page. If there are repetitions in the text, CAT (computer-aided-translation) technology can automatically translate those repetitions.
Most of the texts I get are faxes or poor scans, making CAT statistical analysis impossible. Even good scans are difficult to make a computer text document from.