Add OCR-capability to Apache Solr Attachments

It is a know and already discussed on Drupal.org fact that PDF documents of image type (without text layers) are not searchable. Would be awesome if this module could process files through OCR during upload. Google already uses OCR to search PDF files (http://googleblog.blogspot.com/2008/10/picture-of-thousand-words.html, http://googlesystem.blogspot.com/2008/10/google-uses-ocr-to-index-pdf-fi...) Why this can't be implemented on Drupal?