I've noticed that tineye.com does not attempt to fetch "robots.txt" from my website. It also sends a "user agent" string consisting of the user's browser with "TinEye/1.0 (via http://www.tineye.com/)" tagged on to the end.
Is this correct behaviour? If I write something which fetches one file from a web site, at the user's discretion, is it acceptable to not fetch "robots.txt"? Or should any automated program always obey robots.txt?
Please note this is not a question about tineye.com.