That depends though. Isn't there a way for the PDF page to be an image, but for OCR to produce an index of keywords, so that you are viewing the image, but the index is searchable? This doesn't solve the increase/decrease text size issue (though it does solve the searching issue).