Well, the main thing I was pointing out is that I
know the 900 doesn't treat these sort of errors differently. I occasionally OCR some stuff, and massage the HTML it spits out a bit in an editor, and sometimes my regular expressions screw up and leave a bad tag in somewhere, which I hit when I start reading through. The 505 and 900 seem to behave identically in this situation, and has just told me to add a 'validate XML' step into my process.
